JP2006013917A - Background correction processor, background correction processing method, program for making computer execute the method and image processor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a background correction processor capable of following a two-dimensional fluctuation in background density and performing background elimination that causes little deterioration in image quality without increasing a variation in an image signal of a background part. <P>SOLUTION: In this background correction processor for outputting input image data with the input image data almost uncorrected when the intensity of an image signal of a background part of a target image is equal to that of the image signal on a reference background level, and correcting the image signal of the background level of the target image to the reference background level and outputting an almost black image signal with the image signal almost uncorrected when the intensity of the image is lower, an image data input-output conversion function when the background level C of the target image is lower than the reference background level A is represented by y = ∫(x) when the strength of the image signal on the reference background level is defined as A, the intensity of a black image signal as B, the intensity of the image signal of the background level of the target image as C, input image data as x, micro-displacement as Δ, and output data after background correction processing as y. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a background correction processing apparatus, a background correction processing method, and a program for causing a computer to execute the method for detecting a background level based on an input image signal or image data and performing correction processing according to the detected background level. In particular, the present invention relates to an apparatus that handles image signals or image data generated by reading a document, such as an image scanner, a facsimile, a digital copying machine, and the like.

  There are various types of manuscripts handled by image scanners, facsimiles, digital copiers, etc., and the density of the paper used for the manuscript, that is, the background density, is in a wide range.

  On the other hand, the output image from these devices is desired to have a white background (portion of the paper itself), and if it is output at a density with the background of the original being read, it is stained. As a result, the evaluation of the image quality is lowered.

  For this reason, these devices detect the background level of a document and perform “background correction processing” to make the image data of the background portion constant according to this level, or uniquely detect the background density of the document. Conventionally, “skin removal” is performed to remove portions below the background density.

  For example, in the invention disclosed in Patent Document 1, the black level is assumed to be uncorrected, and an image signal input with an input / output conversion function having a constant slope corresponding to the background level (maximum document reflection luminance) is corrected and output. In the invention disclosed in Patent Document 2, the black side area is equivalent to uncorrected, and the white side area corrects and outputs an image signal input with an input / output function having a constant slope according to the background level.

  That is, the invention disclosed in Patent Document 1 and Patent Document 2 has a property that the slope of the input / output conversion function near the background level becomes steeper as the background level becomes darker.

  By the way, the background density variation in the image data obtained by reading the document includes not only the variation due to the type of the document sheet but also a two-dimensional continuous variation. This is because the background density of the original document itself is not necessarily uniform, the original float due to the folding of the original document or the saddle stitching of the book original document, the influence of flare due to the black solid image existing in the surrounding area, etc. Due to mechanical variation of the mechanism.

  In recent years, there has been an increasing demand for higher image quality, particularly highlight reproducibility of image portions. Conventionally, since the background density cannot be continuously changed in this way, extra background removal is performed. It was necessary and the reproducibility of highlights was insufficient.

In addition, the invention disclosed in Patent Document 3 is intended to prevent show-through and is effective for two-dimensional fluctuation of the background density, but the continuity of the background density is not considered, There is a problem that image quality degradation is severe when an error occurs in the feature determination.
JP-A-6-253135 JP-A-8-307687 JP 2000-209438 A

  As shown in FIG. 25, in practice, the variation in the image signal obtained by reading the background portion of the document does not decrease as the background level becomes darker, but rather tends to increase as the background level becomes darker.

  For this reason, as in the invention disclosed in Patent Document 1, performing correction so that the slope of the input / output conversion function near the background level becomes steeper as the background level becomes darker is contrary to the image signal of the background part. It was supposed to increase the variation.

  Further, as described above, a technique effective for correcting a two-dimensional change in the background density has not been provided conventionally.

The present invention has been made in view of such a problem, and is a background correction processing apparatus, a background correction processing method, a program for causing a computer to execute the method, and an image processing without increasing variations in the image signal of the background portion. An object is to provide an apparatus.
Also provided are a background correction processing apparatus, a background correction processing method, a program for causing a computer to execute the method, and an image processing apparatus that can follow a two-dimensional change in background density and can remove the background with little image quality degradation. The purpose is to do.

  In order to achieve the above object, as a first aspect of the present invention, when the intensity of the image signal of the background portion of the target image is equal to the intensity of the image signal of the reference background level, the input image data is equivalent to uncorrected If the intensity of the image signal of the background portion of the target image is weaker than the intensity of the image signal of the reference background level, the image signal of the background portion of the target image is corrected to the reference background level, A background correction processing apparatus that outputs a substantially black image signal in an uncorrected equivalent state, wherein the intensity of the image signal at the reference background level is A, the intensity of the black image signal is B, and the background portion of the target image When the intensity of the image signal is C, the input image data is x, the minute displacement of the input image data is Δ, and the output data after the background correction processing is y, the background level C of the target image is higher than the reference background level A. Image data when weak When the input / output conversion function y = ∫ (x) is near the intensity C of the image signal of the background portion of the target image, (A−B) / (C−B)> (∫ (C + Δ) −∫ (C)) / A background correction processing apparatus characterized by having the following characteristics: Δ> 0 and (A−B) / (C−B)> (∫ (C) −∫ (C−Δ))> 0 It is. In this way, since the inclination of the image data input / output conversion function is suppressed in the vicinity of the input image data C corresponding to the image background level, the variation in the image data of the background portion can be prevented from being enlarged.

In the first aspect of the present invention, the image data input / output conversion function y = ∫ (x) is
1 = (∫ (c + Δ) −∫ (C)) / Δ and 1 = (∫ (C) −∫ (C−Δ)) / Δ
It is preferable to have the following characteristics. In this way, since the inclination related to the image data input / output conversion is 1 in the vicinity of C of the input image data corresponding to the image background level, it is possible to further suppress variations in the image data of the background portion. Or, the image data input / output conversion function y = ∫ (x) is
1> (∫ (c + Δ) −∫ (C)) / Δ and 1> (∫ (C) −∫ (C−Δ)) / Δ
It is preferable to have the following characteristics. In this way, since the slope of the image data input / output conversion function is less than 1 in the vicinity of the input image data C corresponding to the image background level, it is possible to suppress variations in the image data of the background portion.

  In order to achieve the above object, as a second aspect of the present invention, when the intensity of the image signal of the background portion of the target image is equal to the intensity of the image signal of the reference background level, the input image data is not corrected. Output in the same state, and if the intensity of the image signal of the background portion of the target image is weaker than the intensity of the image signal of the reference background level, the image signal of the background portion of the target image is corrected to the reference background level. A background correction processing method for outputting a substantially black image signal in an uncorrected equivalent state, wherein the intensity of the image signal at the reference background level is A, the intensity of the black image signal is B, and the background of the target image When the intensity of the image signal of the portion is C, the input image data is x, the minute displacement of the input image data is Δ, and the output data after the background correction processing is y, the background level C of the target image is the reference background level A. Image when weaker than When the data input / output conversion function y = ∫ (x) is in the vicinity of the intensity C of the image signal of the background portion of the target image, (AB) / (CB)> (∫ (C + Δ) −∫ (C)) A background correction processing method characterized by having the following characteristics: / Δ> 0 and (A−B) / (C−B)> (∫ (C) −∫ (C−Δ)) / Δ> 0 It is to provide. In this way, since the inclination of the input / output conversion function of the image signal is suppressed in the vicinity of the input image signal C corresponding to the image background level, the variation in the image signal of the background portion can be prevented from being enlarged.

In the second aspect of the present invention, the image data input / output conversion function y = ∫ (x) is
1 = (∫ (c + Δ) −∫ (C)) / Δ and 1 = (∫ (C) −∫ (C−Δ)) / Δ
It is preferable to have the following characteristics. In this way, since the slope of the input / output conversion function of the image signal is 1 in the vicinity of the input image signal C corresponding to the image background level, it is possible to further suppress the spread of the image signal variation in the background portion. Or, the image data input / output conversion function y = ∫ (x) is
1> (∫ (c + Δ) −∫ (C)) / Δ and 1> (∫ (C) −∫ (C−Δ)) / Δ
It is preferable to have the following characteristics. In this way, since the slope of the input / output conversion function of the image signal is less than 1 in the vicinity of the input image signal C corresponding to the image background level, variations in the image signal of the background portion can be suppressed.

  Moreover, in order to achieve the said objective, this invention provides the program which makes a substantial computer perform the background correction | amendment processing method concerning the structure in any one of the said 2nd aspect of the said invention as a 3rd aspect. Is.

  In order to achieve the above object, the present invention provides, as a fourth aspect, a representative background level that is a background level that represents the entire image based on image data obtained by performing main / sub-scanning on the target image. Based on the first background detection means for detecting, the second background detection means for detecting the local background level, which is the local background level of the target image, based on the detected representative background level, and the local background level An image processing apparatus having a background correction unit for correcting image data, wherein the second background detection unit stores a background level in each section when the main scanning direction of the image data is divided into a plurality of sections. And a background level predicting unit that predicts the background level of the target section based on the background level around the target section stored in the storage unit when any of the sections is the target section. The local background level is determined on the basis of the extraction means for extracting the feature amount of the image data of the target section, and the predicted background level and the background level. An image processing apparatus comprising: a determination unit, wherein the storage unit stores a representative background level as an initial value, and stores a local background level after the start of main / sub scanning. It is to provide. In this way, image quality degradation at the sub-scanning front end is small, and image quality degradation when an error occurs in feature determination can be reduced.

  In the fourth aspect of the present invention, it is preferable that the predicting means predicts the background level by performing a weighted average of the background levels in the section around the section of interest stored in the storage means. In this way, it is possible to reduce image quality degradation when an error occurs in feature determination.

  In any of the above-described configurations of the fourth aspect of the present invention, the extracting means includes means for extracting the undulations of light and dark as the feature amount of the image data of the target section, and the determining means is the light and dark extracted by the extracting means. The background processing means for determining whether or not the image of the target section is considered to be the background by the threshold processing for the undulation, and the background level predicted by the prediction means when the background determination means determines that the image is not considered the background. It is preferable to provide a means for determining the background level. In this way, it is possible to maintain the accuracy of the detected background data determined as the background level of the target section, that is, the character portion and the halftone portion are determined to be the background by checking the shading of light and dark. It is possible to prevent the fluctuation of the detected background level by extracting a feature amount that directly represents the feature of the background portion. In addition, by checking that the brightness is above a certain level, the detected background level does not become too dark, so it is possible to process a document whose image density changes continuously, and by checking that the color is achromatic The accuracy of background determination can be increased.

  In any of the above configurations of the fourth aspect of the present invention, the extracting means includes means for extracting saturation as the feature amount of the image data of the target section, and the determining means is the saturation extracted by the extracting means. In response, the background determination means for determining whether or not the image of the target section is considered to be the background, and the background level predicted by the prediction means when the background determination means determines that the background determination means does not regard the background as the local background level. And means for determining. In this way, it is possible to maintain the accuracy of the detected background data determined as the background level of the target section, that is, the accuracy of the background determination can be increased by checking that the background color is achromatic. .

  In any of the above configurations of the fourth aspect of the present invention, the extraction means includes means for extracting average brightness as the feature amount of the image data of the target section, and the determination means is extracted by the extraction means. The background determination means for determining whether or not the image of the target section is regarded as the background by threshold processing for the average brightness, and the background level predicted by the prediction means when the background determination means determines that the image is not considered as the background. It is preferable to provide a means for determining as a local background level. In this way, it is possible to maintain the accuracy of the detected background data determined as the background level of the target section, that is, the detected background level becomes darker by checking that the brightness is above a certain level. Since it does not become too much, it is possible to process an original whose image density changes continuously.

  In the fourth aspect of the present invention, in the configuration in which the extraction means includes means for extracting at least one of light undulation, saturation, and average brightness, the extraction means is a feature of the image data of the target section. An image average value calculating unit that calculates an average value of the image data of the target section as a quantity, and the determining unit calculates a weighted average of the background level predicted by the predicting unit and the average value calculated by the image average value calculating unit. It is more preferable to include a weighted average background level calculating unit to calculate, and a unit to determine the background level calculated by the background level calculating unit as a local background level when the difference determining unit determines that the background can be regarded as a background. In this way, it is possible to prevent the detected background level from changing suddenly, and to reduce processing unevenness due to the background correction processing.

According to the present invention, it is possible to provide a background correction processing apparatus, a background correction processing method, a program for causing a computer to execute the method, and an image processing apparatus that do not increase variations in the image signal of the background portion.
Also provided are a background correction processing apparatus, a background correction processing method, a program for causing a computer to execute the method, and an image processing apparatus that can follow a two-dimensional change in background density and can remove the background with little image quality degradation. it can.

[Principle of the Invention]
The background correction processing method according to the present invention suppresses the slope of the input / output conversion function near the background level when the background level is dark.

  FIG. 1 shows an example of the input / output conversion function. In FIG. 8, A represents a standard background level, and B represents a black level. Here, if the background level of the target image is equal to the standard background level A, a straight line connecting coordinates (A, A) and coordinates (B, B) is used as an input / output conversion function. In this case, the input image signal is output in a state equivalent to uncorrected.

  On the other hand, when the background level of the target image is darker than the standard background level A as in C, a straight line passes through the coordinates (C, A) and connects the coordinates (C, A) and the coordinates (B, B). An input / output conversion function having a gentler slope before and after the coordinates (C, A) than the slope (AB) / (CB) is used.

  That is, if the input / output conversion function is y = ∫ (x) and the minute displacement of the input image signal is Δ, the characteristic shown in the equation (1) is obtained.

  In Equation (1), the condition that the slope in the vicinity of the coordinates (C, A) is greater than 0 is that the smoothing process is performed after the background correction process. That is, if one is limited to 0, that is, the limiter is applied only to one side, only the variation on the opposite side remains, and even if the smoothing process is applied to this result, the original background level is not obtained.

  Further, in this case, since the variation in the image signal after the correction of the background portion is more gradual, the conventional correction (broken line) using a straight line connecting the coordinates (C, A) and the coordinates (B, B) or the like. Smaller (solid line). That is, according to the background correction processing method using the input / output conversion function as shown in Expression (1), it is possible to prevent the variation in the image signal of the background portion from being enlarged.

  In the input / output conversion function having the above characteristics, for example, the coordinates of the maximum value of the input / output image signal are (D, D), and the above coordinates (C, A) and coordinates (B, B) are used. It can be obtained by calculating an arc passing through the added three points or a quadratic function.

For example, in the latter case, the input / output conversion function is expressed by y = ∫ (x) = α · x ² + β · x + γ (2)
And solving the ternary linear simultaneous equation shown in equation (3), the input / output conversion function of equation (2) can be obtained.

Next, FIG. 2A shows a configuration of an image processing apparatus including a background correction processing apparatus that performs the background correction processing method. The image processing apparatus includes an image signal generation apparatus 901, a background detection processing apparatus 902, a background correction processing apparatus 903, an image signal reception apparatus 906, and a system control apparatus 907.
The image signal generation device 901 is a functional unit that generates an image signal of a target image, and is an image signal generation device such as a known flatbed scanner. The image signal generated by the image signal generation device 901 is input to the background detection processing device 902, and the background level of the target image is detected. A known configuration can be provided for the background detection processing device 902. The background level detected by the background detection device 902 is input to the calculation unit 904 of the background correction processing device 903.

  The calculation unit 904 is configured by a general microcomputer or the like, and solves a ternary simultaneous equation as shown in the equation (3), for example, according to the detected background level, and performs an input / output conversion function. To decide. Next, an input / output conversion table is generated based on this and set in the conversion unit 905.

  The conversion unit 905 is a so-called look-up table (LUT), and converts the image signal generated by the image signal generator 901 into a table. That is, when the input / output conversion table is set in the conversion unit 905, the image signal generation unit 901 outputs the image signal of the target image again, and the conversion unit 905 converts the table-converted image signal into an image signal receiving device. Output to 906.

  The image signal receiving device 906 is a printer that records an image based on the image signal on paper or the like, for example, and receives the image signal output from the background correction processing device 903.

  The image signal generation device 901, the background detection processing device 902, the background correction processing device 903, and the image signal processing device 906 are connected to the system control device 907. The system control device 907 controls and links each device.

  When the background level of the target image is darker than the standard background level, the background correction device 903 passes the coordinates (C, A) and the coordinates (B, B) in FIG. Since the background correction processing is performed using the input / output conversion function before and after the inclination coordinates (C, A) which is gentler than the inclination (AB) / (CB) of the connecting line, the image signal of the background portion is The variation is not enlarged.

FIG. 2B shows a configuration of another image processing apparatus provided with a background correction processing apparatus. The image processing apparatus illustrated in FIG. 2B includes an image signal generation apparatus 911, a background detection processing apparatus 912, a background supplement processing apparatus 913, an image signal reception apparatus 916, and a system control apparatus 917.
The image signal generation device 911, the background detection processing device 912, and the image signal reception device 916 are the same as those having the same names in the configuration shown in FIG.

  The background correction processing device 913 is a so-called two-dimensional LUT, and converts the image signal generated by the image signal generation device 911 into a table according to the detected background level input from the background detection processing device 912. Note that this two-dimensional LUT solves, for example, a ternary simultaneous equation as shown in Equation (3), generates input / output conversion tables for the background level where detection is predicted, and collects them. This can be realized by using a ROM for storing.

  The image signal generation device 911, the background detection processing device 912, the background correction processing device 913, and the image signal reception device 916 are connected to the system control device 917, and the system control device 917 controls and interlocks each unit.

  When the background level of the target image is darker than the standard background level, the background correction device 913 passes the coordinates (C, A) and coordinates (C, A) and (B, B) in FIG. Since the background correction processing is performed using the input / output conversion function before and after the inclination coordinates (C, A) which is gentler than the inclination (AB) / (CB) of the connecting line, the image signal of the background portion is The variation is not enlarged.

Next, another example of the input / output conversion function when the background level of the target image is darker than the standard background is shown in FIG. In FIG. 3, A is a standard background level, B is a black level, C is a background level of a target image, and D is a maximum value of an input / output image signal.
In this input / output conversion function, a width w in consideration of variations in the image signal of the background portion is determined in advance, and a straight line connecting the coordinates (B, B) and the coordinates (D, D) and an inclination equal to this straight line. (A slope 1) and a straight line passing through the coordinates (C, A).

  That is, in terms of the slope of the input / output conversion function, the slope is 1 when the level of the input image signal is B or in the region between (Cw), and in the range of the input image signals B to CW. , Both ends are divided into two linear parts (1001 and 1002) which are continuously changed from 1 as a starting point. Such an input / output conversion function is close to a straight line connecting the coordinates (B, B) and the coordinates (D, D) near the black level, and the coordinates (( It gradually approaches a straight line passing through C, A), and in a region where the input image signal is Cw or more, it has a slope of 1 and coincides with a straight line passing through coordinates (C, A). However, the maximum value of the output image signal is limited by D.

The slope of this input / output conversion function is 1 in the vicinity of the input signal C corresponding to the image background level. That is, if the input / output conversion function is y = ∫ (x) and the minute displacement of the input image signal is Δ,
1 = (∫ (C + Δ) −∫ (C)) / Δ
And 1 = (∫ (C) −∫ (C−Δ)) / Δ
It becomes. Therefore, the background correction process can be performed without increasing the variation of the image signal of the background portion.

  Note that the slope in the range of the input image signals B to C-w does not need to be divided into two linear portions as long as it changes continuously from 1 at both ends. For example, it may be divided into three straight portions 1003, 1004, 1005 as shown by dotted lines in FIG. 3 or more, or a curve such as a cosine curve 1006 shown by broken lines in FIG.

  The reason why the slope in the region equal to or greater than Cw is 1 is to prevent the input / output conversion function from changing suddenly when the background level of the target image approaches the standard background level. That is, when the background level of the target image approaches the standard background level as much as possible, the input / output conversion function shown in FIG. 3 approaches the straight line connecting the coordinates (B, B) and the coordinates (D, D) as much as possible. . Accordingly, in this case as well, the input image signal is output in a state equivalent to uncorrected.

Also. FIG. 4 shows another example of the input / output conversion function when the background level of the target image is darker than the standard background level. In FIG. 4, A is a standard background level, B is a black level, C is a background level of a target image, and D is a maximum value of an input / output image signal.
Also in this input / output conversion function, a width w in consideration of variations in the image signal of the background portion is determined in advance, and a straight line connecting coordinates (B, B) and coordinates (D, D), and coordinates (C, A straight line connecting A) and coordinates (D, D) is synthesized.

In terms of the slope of the input / output conversion function, the slope is 1 when the input image signal is B, and the slope is (DA) / (DC) in the region of C-w or higher. Since C <A, (DA) / (DC) is less than 1.
Further, in the range of the input image signals B to Cw, the inclination is continuously such that the inclination of the end on the B side is 1 and the inclination on the Cw side is (DA) / (DC). Change. For example, as shown in FIG. 4, the amount of change in inclination changes as represented by two linear portions 1101 and 1102.

  Such an input / output conversion function is close to a straight line connecting the coordinates (B, B) and the coordinates (D, D) in the vicinity of the black level, and is gradually coordinated (C, A) as the input image signal approaches Cw. Near the straight line connecting the coordinates (D, D) and the area where the input image signal is C-w or more, it matches the straight line connecting the coordinates (C, A) and the coordinates (D, D).

Thus, the slope of the input / output conversion function shown in FIG. 4 is less than 1 near the input image signal C corresponding to the image background level. That is, if the input / output conversion function is y = ∫ (x) and the minute displacement of the input image signal is Δ,
1> (∫ (C + Δ) −∫ (C)) / Δ
And 1> (∫ (C) −∫ (C−Δ)) / Δ
It becomes. Therefore, it is possible to perform background correction processing while reducing variations in the image signal of the background portion.

  Further, it has been found that the variation in the image signal obtained by reading the background portion of the document tends to increase as the background level becomes darker. On the other hand, since the input / output conversion function shown in FIG. 4 has an inclination of (DA) / (DC) in the vicinity of the input image signal C corresponding to the image background level, the inclination becomes smaller as the background level becomes darker. Become. Therefore, it can be suppressed that the variation in the image signal becomes larger as the background level becomes darker.

In the above description, the case where the slope of the input / output conversion function changes continuously is taken as an example, but the present invention is not limited to this.
For example, in the example of the input / output conversion function shown in FIG. 5, A is the standard background level, B is the black level, C is the background level of the target image, D is the maximum value of the input / output image, and w is the image of the background portion. This is a width that takes into account the variation of. The input / output conversion function includes a straight line passing through coordinates (C, A) and having a slope of 1, and coordinates (Cw, Aw) and coordinates U (B , B). The maximum value of the output image signal is regulated by D.

  This input / output conversion function is an example in which the input / output conversion function shown in FIG. 3 is approximated by a broken line, and has the same effect.

  The input / output conversion function shown in FIG. 6 is an example in which the input / output conversion function shown in FIG. 4 is approximated by a broken line, and connects the coordinates (B, B) and the coordinates (Cw, Aw). It consists of a straight line and a straight line connecting coordinates (Cw, Aw) and coordinates (D, D). This input / output conversion function has the same effect as the input / output conversion function shown in FIG.

  Also, the input / output conversion function shown in FIG. 7 is another example in which the input / output conversion function shown in FIG. In FIG. 7, A is the standard background level, B is the black level, C is the background level of the target image, D is the maximum value of the input / output image signal, and w is the width considering the variation of the image signal of the background portion. A range where the inclination is 1 is provided on the black level side. E is this upper limit.

  That is, the input / output conversion function is a straight line having a slope of 1 passing through the coordinates (C, A), and the coordinates (Cw, Aw) and coordinates where the input image signal is a point Cw on the straight line. (E, E) and a straight line connecting coordinates (E, E) and coordinates (B, B). However, in this input / output conversion function, the maximum value of the output image signal is limited by D.

  The input / output conversion function shown in FIG. 8 is another example in which the input / output conversion function shown in FIG. 4 is approximated by a polygonal line, and a straight line connecting coordinates (B, B) and coordinates (E, E), and coordinates ( E, E) and a straight line connecting coordinates (Cw, Aw), and a straight line connecting coordinates (Cw, Aw) and coordinates (D, D). This input / output conversion function has the same effect as the input / output conversion function shown in FIG.

  In the above description, the width w considering the variation of the image signal in the background portion is described as being fixed, but is not limited thereto. That is, it has been found that the variation in the image signal obtained by reading the background portion of the document tends to increase as the background level becomes darker. Therefore, if the width w is changed according to the image background level. Since it is possible to control the range of the gradient 1 and the gradient (DA) / (DC) having a variation suppressing effect, it is possible to cope with such a tendency.

  Since it is obvious that the background correction processing method shown in FIGS. 3 to 8 can be executed by the background correction processing apparatuses 903 and 913 shown in FIGS. 2A and 2B, description thereof will be omitted.

  Here, an outline of the background detection method according to the present invention will be described. FIG. 9 is an explanatory diagram of this background detection method. In FIG. 9, a document 301 is set within an image data range 302. Note that the image data indicates data generated by optically reading the image data range 302 into pixels. The image data is basically premised on data read by the raster scan method in the main and sub directions. In FIG. 9, the horizontal direction of the paper is the main scan direction of the raster scan, and the vertical direction of the paper is the sub scan direction. .

The feature of this background detection method is that the main scanning direction is divided into a plurality of sections and handled, and the background is detected for each section. For this purpose, the digital image copying apparatus according to the present embodiment includes a background level buffer 303 that holds the background level for each section.
Here, the background detection processing operation is performed on the image data input in accordance with this reading method. Therefore, if the image signal at the sub-scanning position 304 is to be processed now, the background level buffer 303 is determined by the immediately preceding main scanning line (main scanning line at the immediately preceding sub-scanning position). The background level is maintained.

  Further, in order to detect the background for each section, attention is focused on a section 305 in the main scanning direction at the sub-scanning position 304 as a detection section (hereinafter, this section is also referred to as “attention area” or “attention section”). Usually, the background level in the image data hardly varies depending on the position, and even if it changes, it is gentle. Therefore, the background level of the attention area can be predicted by referring to the background level of the adjacent peripheral area, and the background level prediction processing unit 306 stores the background level in the background level buffer 303 as a peripheral section adjacent to the attention area. The background level of the attention area is predicted with reference to the section corresponding to the attention area (in other words, determined in the immediately preceding main scan) and the background levels of the left and right adjacent sections.

The predicted value at the background level may be directly applied as the background level of the region of interest when there is very little variation due to the position. Correct and use this as the appropriate background level.
Note that when the attention area is not the background, correction cannot be performed, and thus the predicted value is applied as the background level of the attention area.

This background detection processing is executed by referring to the image data of the attention area to determine whether or not the attention area is a background portion, and if it is a background portion, the predicted background level is corrected based on the image data. To do. In FIG. 9, the feature amount extraction unit 307 refers to the image data of the attention area 305 and extracts a feature amount used for determining whether or not it is a background portion.
The background level determination processing unit 308 determines whether or not the background portion is a background portion based on the feature amount extracted by the feature amount extraction unit 307. If it is determined that the background portion is the background portion, the predicted background level is determined by the feature amount. If the background level of the attention area 305 is corrected and determined not to be the background portion, the predicted background level is determined as it is as the background level of the attention area 305.

  The feature amount extraction and the background level determination process using the feature amount will be described in detail later.

The background level determined by the background level determination unit 308 is stored in the background level buffer 303 as detection data of the attention section 305, and is used as data of a peripheral section adjacent to the attention area in the background level determination processing in the next line.
In the main scanning (first line) at the first sub-scanning position, the background level buffer 303 does not store the background level in the previous main scanning, but the representative background level of the document is detected in advance. This is stored as an initial value. As a result, the background detection process can be executed from the first line in the sub-scanning direction.

  As described above, this background detection method can detect an appropriate background level for each divided section in accordance with the progress of the raster scan and also following the two-dimensional background level fluctuation.

  As described above, the background detection method can sequentially detect the background level without performing pre-scanning, and can follow the fluctuation of the two-dimensional background level.

[Mechanical structure]
A digital copying machine that preferably implements the present invention will be described. FIG. 10 shows a schematic configuration of the mechanism part of the digital copying machine. The mechanism of the digital copying machine can be broadly divided into a scanner unit 101 that reads a document and a printer unit 102 that records an image on a recording sheet.

  The original 103 is placed at a predetermined position on the platen 104 and illuminated by the halogen lamps 105a and 105b. Reflected light from the original 103 passes through the first mirror 106, the second mirror 107, the third mirror 108, and the lens 109, forms an image on the CCD 110 that is a three-line color line image sensor, and is photoelectrically converted into an image signal by the CCD 110. . Here, the halogen lamps 105a and 105b and the first mirror 106 are mounted on a first carriage (not shown), and the second mirror 107 and the third mirror 108 are mounted on a second carriage (not shown). A carriage drive motor (not shown) moves the first carriage and the second carriage at a speed ratio of 2: 1. As a result, the entire surface of the document placed on the platen 104 is scanned while keeping the optical path length between the document and the lens constant.

  The image signal photoelectrically converted by the CCD 110 is subjected to various processes by the image processing unit 111 and the like, and then input to an LD (laser diode) (not shown) of the printer unit 102, where it is converted into laser light.

  The laser beam emitted from the LD (not shown) is reflected by the polygon mirror 112, passes through the fθ lens 113 and the fourth mirror 114, and is imaged and illuminated on the surface of the photosensitive drum 115 rotating counterclockwise. The Here, the polygon mirror 112 is fixed to the rotating shaft of the polygon motor 116, and the polygon motor 116 rotates at a constant speed to rotate the polygon mirror 112. Further, by the rotation of the polygon mirror 112, the laser beam is scanned in a direction perpendicular to the rotational movement direction of the photosensitive drum 115, that is, a direction along the drum axis.

  On the other hand, the surface of the photosensitive drum 115 is charged to a uniform positive potential in advance by a charging charger 117 connected to a high voltage generator (not shown). When the surface of the photosensitive drum 115 is irradiated with laser light, the surface charge flows to the equipment ground of the drum body due to a photoconductive phenomenon and disappears. Here, the portion where the document density is light is turned on with a weak LD, and the portion where the document density is dark is turned on with a strong LD. As a result, an electrostatic latent image corresponding to the density of the document is formed on the surface of the photosensitive drum 115 by main scanning by the polygon mirror 112 and sub scanning by rotation of the photosensitive drum 115.

  The document unit 118 includes developing units K, C, M, and Y that respectively store black, cyan, magenta, and yellow positively charged toners, and any one of the developing units is selected. The selected developing unit is biased to a predetermined positive potential by a high voltage generator (not shown), develops the electrostatic latent image, and forms a toner image corresponding to the density of the document on the surface of the photosensitive drum 115. .

  The transfer belt 119 is biased to a predetermined negative potential by a high voltage generator (not shown), and rotates clockwise at the same speed as the photosensitive drum 115. The toner image is attracted by the action of a bias and transferred onto the surface of the transfer belt 119 while the photosensitive drum 115 and the transfer belt 119 are close to each other.

  The electrostatic latent image formation, toner image formation, and toner image transfer operation are repeated as many times as necessary. That is, in each of the full color, document color, and registered color modes, the process is performed four times, and the developing unit is selected in the order of K, C, M, and Y, and the formed toner image is aligned. Thus, the image is transferred onto the surface of the transfer belt 119. In the case of each mode of black, cyan, magenta, and yellow, the process is performed only once, and the developing portions K, C, M, and Y are selected, respectively. In each of the red, green, and blue modes, the process was performed twice, and each of the developed portions M → Y, C → Y, or C → M was selected, and the formed toner image was aligned. The image is transferred onto the surface of the transfer belt 119.

  On the other hand, recording sheets 121a and 121b are respectively stored in the sheet feeding cassettes 120a and 120b, and one of the sheet feeding cassettes is selected. The recording paper 121a of the selected paper feeding cassette, for example, the paper feeding cassette 120a is fed out by the paper feeding operation of the paper feeding roller 122a and reaches the registration rollers 123a and 123b. The registration rollers 123a and 123b are initially stopped and start rotating at a predetermined timing in accordance with the position of the toner image on the rotating transfer belt 119 to feed out the recording paper.

  The transfer charger 124 is connected to a high voltage generator (not shown) that generates a negative voltage, and the toner image on the transfer belt 119 is retransferred to the recording paper being fed by the action of the transfer charger 124. . When the toner image is retransferred to the recording paper, the bias of the transfer belt 119 is released to promote the retransfer.

  The recording paper onto which the toner image has been retransferred is sent to the heat fixing units 125a and 125b, and the toner image is fixed thereto. The recording paper to which the toner image is fixed is discharged out of the apparatus.

  The toner remaining on the surface of the photosensitive drum 115 after the transfer is removed by the cleaning unit 126, and the photosensitive drum 115 is ready for the next operation. Further, the toner remaining on the surface of the transfer belt 119 after the retransfer is removed by the cleaning unit 127, and the transfer belt 119 is ready for the next operation.

[Electrical configuration]
FIG. 11 shows the configuration of the electrical system of the digital copying machine 202.
The electrical component of the digital copying machine 202 is divided into a scanner unit 101, an image processing unit 111, a printer unit 102, an operation display unit 201, and a system control unit 203. The scanner unit 101 and the printer unit 102 have been described above. The operation display unit 201 detects input operations such as processing mode selection and displays information. The system control unit 203 communicates with the control circuit of each unit and controls the overall operation of the digital copying machine 202.

  The digital copying machine 202 is connected to a LAN connection device 204, and inputs / outputs image signals to / from other devices such as a workstation 206 and an information processing terminal 207 connected to the LAN 205 via the LAN connection device 204. To do.

  FIG. 12 shows the configuration of the scanner unit 101. The CCD 110 color-separates incident light into red, green, and blue, and then performs photoelectric conversion to output three types of analog image signals to the A / D conversion circuit 208. The A / D conversion circuit 208 converts the input image signal into a digital signal while performing correction of the dark current of the CCD 110, and outputs the digital signal to the shading correction circuit 209.

The shading correction circuit 209 is a circuit that corrects uneven illumination of the halogen lamps 105a and 105b, uneven sensitivity of the light receiving elements in the CCD 110, and the A / D conversion circuit when the white reference plate 126 shown in FIG. 10 is read. 208 output signals are stored, and based on this, the gain of the image signal for each color and each pixel is adjusted. Thereby, for example, the output image signal of the shading correction circuit 209 when the white reference plate 126 is read becomes a predetermined value.
Note that the shading correction circuit 209 also performs processing such as correction for a reading position shift between red, green, and blue image signals due to the three-line structure of the CCD 110.

  The scanner control circuit 210 communicates with the system control unit 203 and controls the entire scanner unit 101. For example, the operation control of the A / D conversion circuit 208 and the shading correction circuit 209, and the halogen lamps 105a and 105b. Lighting control and rotation control of the carriage drive motor 211 are performed. The document size sensor 212 is a sensor that detects the size of the document placed on the platen, and outputs the detection result to the scanner control circuit 210.

  On the other hand, the image signal output from the shading correction circuit 209 is input to the background removal circuit 213 of the image processing unit 111.

  FIG. 13 shows the configuration of the image processing unit 111. The background detection circuit 213 sequentially detects the background level of the document based on the image signal input from the shading correction circuit 209, and outputs the image signal corrected in accordance with the background level to the γ conversion circuit 214.

  The γ conversion circuit 214 is connected to the LAN connection device 204 in addition to the background removal circuit 213, and gradation of the image signal input from the background removal circuit 213 or the LAN connection device 204 into an image signal proportional to the density or the like. The image is converted and output to the image area separation circuit 215, the area control circuit 216, and the delay circuit 217. Further, depending on the setting, the γ conversion circuit 214 may perform gradation conversion on the image signal input from the background removal circuit 213 and output the image signal to the LAN connection device 204.

  The image area separation circuit 215 is a circuit that determines whether the image portion being processed is a line drawing such as a character, a monochrome image, or a color image based on the input image signal, and a determination signal that represents the result Is output to the filter circuit 218.

  The delay circuit 217 delays the determination of the image area separation circuit 215, and therefore delays the image signal in accordance with this delay and outputs it to the filter circuit 218.

  The area control circuit 216 is a circuit for storing an input image signal, generating a switching signal for performing different image processing for each partial area of the image, and the like. Output to.

  The filter circuit 218 is a circuit that outputs the image signal from the delay circuit 217 by performing two-dimensional filter processing such as edge enhancement and smoothing. These processes are controlled by a determination signal or a switching signal. For example, if the determination signal is a line drawing, edge enhancement filter processing is performed, and if the determination signal is a non-line drawing, smoothing filter processing is performed. Further, edge enhancement or smoothing filtering may be performed by the switching signal regardless of the determination signal.

  The color correction circuit 219 performs color correction processing for converting the image signal from the filter circuit 218 into a toner recording amount corresponding to the developing device selected as described above. This processing is controlled by a determination signal and a switching signal.For example, the toner recording amount suitable for the full color mode or the black mode is converted according to the switching signal, or the determination signal is a line drawing and a monochrome image in the full color mode. For example, conversion may be performed in which the UCR rate (Under Color Removal) is set to 100%, and the UCR rate is set to 70% for a non-line image or a color image. In some cases, a filling process for outputting a constant image signal value and a determination signal value according to the switching signal and a color conversion process for converting a designated color into another color may be performed. From the color correction circuit 219, one type of image signal converted into the toner recording amount, a determination signal, and a switching signal are output and input to the scaling circuit 220.

  The scaling circuit 220 is a circuit that performs scaling processing for enlarging / reducing the three types of signals from the color correction circuit 219 in the main scanning direction. Note that the scaling process for enlarging / reducing the document in the sub-scanning direction is performed by controlling the rotation of the carriage drive motor 211.

  The gradation processing circuit 221 performs γ correction processing for correcting the recording characteristics (image signal versus toner recording amount characteristics) of the printer unit 102 that fluctuate depending on temperature and humidity, and 1 or several pixels in the main scanning direction and 1 in the sub-scanning direction. Alternatively, gradation processing or the like is performed with several pixels as one unit and a stretched expression that emphasizes resolution or gradation. Here, the γ correction processing and the gradation processing are controlled by a determination signal or a switching signal. For example, if the determination signal is a line drawing, gradation processing is performed with an emphasis on resolution, and if the determination signal is a non-line drawing, the gradation processing is performed. Gradation processing with emphasis on performance is performed. Further, predetermined gradation processing or γ correction processing may be performed by the switching signal regardless of the determination signal.

  Note that the gradation processing circuit 221 is also connected to the LAN connection device 204, and can selectively output an image signal subjected to the above processing or an image signal input from the LAN connection device 204.

  The image processing control circuit 222 communicates with the system control unit 203 and, in response to a request from the system control unit 203, a background removal circuit 213, a γ conversion circuit 214, an image area separation circuit 215, an area control circuit 216, a delay Settings of the circuit 217, the filter circuit 218, the color correction circuit 219, the scaling circuit 220, and the gradation processing circuit 221 are performed to control the entire image processing unit 111.

  The image signal output from the gradation processing circuit 222 is input to the LD control circuit 223 of the printer unit 102.

  FIG. 14 shows the configuration of the printer unit 102. The LD control circuit 223 performs pulse width modulation and power modulation according to the input image signal and controls the entire printer unit 102. For example, the LD control circuit 223 is controlled to forcibly turn off the LD 224, the rotation control of the polygon motor 116, the motor 226 that rotationally drives the photosensitive drum, and the motor 227 that rotationally drives the transfer belt, and development of the developing unit 118. Selection control of the devices K, C, M and Y, output control for each load of the high voltage power supply 228 (for example, charging charger, developing device, transfer belt, transfer charger, etc.), temperature control of the thermal fixing units 125a, 125b, etc. Do. The printer unit 102 has a paper size sensor 229 for detecting the size of the stored recording paper for each of the paper feed cassettes 120 a and 120 b, and the detection result is output to the printer control circuit 225.

  FIG. 15 shows the configuration of the operation display unit 201. The operation display unit 201 includes a touch panel display (TPD) 230 in which a display unit that displays various mode options, setting states, read document images, and the like and a detection unit that detects a pressed position of the display unit are integrated. And a keyboard 231 for inputting the number of copies, a copy start command, and the like, and an operation display control circuit 232. The operation display control circuit 232 prompts key input by displaying options for each mode on the TPD 230, detects input from the TPD 230 and the keyboard 231 and displays a state set in the TPD 230, and also displays a system control unit. 203 communicates with the terminal 203 to transmit the input result. Also, the TPD 230 is controlled when the read image signal of the document is sent to the TPD 230 via the area control circuit 216.

[First Embodiment]
FIG. 16 shows a detailed configuration of the background removal circuit 213 that executes the background removal processing according to the present invention.
The background removal circuit 213 includes a background level detection circuit 401 that detects the background level of the document image based on the input color image signal, and a background level correction circuit 402 that corrects the image signal according to the detected background level. Has been.

  The background level storage circuit 403 is a circuit that stores the color background level for each section when the main scanning direction of the image signal is divided into a plurality of sections, and the main scanning of the image signal input to the background detection circuit 401. The stored background level is output to the background level prediction circuit 404 according to the direction position.

  The background level prediction circuit 404 corresponds to a peripheral section of a section (attention section) that is a target of processing by a representative value extraction circuit 405 and a flatness determination circuit 406, which will be described later, and is stored in the background level device large depth circuit 403. It is a circuit that predicts the background level of the target section based at least on the level, and outputs the predicted background level to the light / dark change evaluation circuit 409, the background portion determination circuit 410, and the background level determination circuit 411. As shown in Equation (4), the background level prediction is realized by calculating a weighted average of the background level stored in the target section and its right and left adjacent sections for each color component, and setting this as the predicted background level. To do.

  Predicted background level = (8 × (stored background level of the section of interest) + 12 × stored background level of the next left section + 12 × (stored background level of the right next section)) / 32 (4)

  On the other hand, the input color image signal is input to the representative value extraction circuit 405 and the flatness determination circuit 406.

  The representative value extraction circuit 405 is a circuit that extracts and outputs an image signal value (representative value) representing a corresponding section, that is, a section of interest, for each color component in accordance with the position of the input image signal in the main scanning direction. is there. The representative value extraction circuit 405 outputs the extracted representative value to the achromatic color determination circuit 407, the brightness determination circuit 408, the light / dark change evaluation circuit 409, and the background portion determination circuit 410. The representative value is extracted by calculating the average value or the intermediate value of the image signals in the target section (the signal value positioned in the middle when arranged in the order of the size of the image signal), or the average value or the intermediate value of the image signals in the target section and its surroundings. This is realized by determining this as a representative value. For example, when the representative value is the average value of the image signals in the target section, the calculation is performed as shown in Expression (5).

  Representative value = Σ (total sum of image signals in the target section) / (number of pixels in the target section) (5)

  The flatness determination circuit 406 is a circuit that determines the flatness of the image signal in the attention section, and outputs the flatness determination result to the background portion determination circuit 410. The determination of flatness is realized by, for example, extracting the maximum value and the minimum value of the image signal in the target section for each color component, calculating the difference, and determining that each is flat if it is smaller than a predetermined threshold value. To do.

  The flatness determination result is used to determine the background portion as will be described later. That is, the image signal of the character / thin line portion or halftone print portion of the document has a feature that the shading of the light and dark is severe, and the flatness determination circuit 406 does not determine such a portion as the background portion. The presence or absence of undulations in the image signal is determined by flatness.

  Further, the variation amount of the image signal in the actual background portion varies depending on the color component. This occurs, for example, when the signal S / N ratio at the time of reading a document differs depending on the color component. In equation (6), a constant that is different for each color component is used as a threshold value. This is because the variation of each color component is taken into account, and this improves the accuracy of determination of flatness.

  Note that the flatness determination may be performed based on the image signal of the section of interest and its surroundings. Alternatively, a statistic such as the variance of the image signal in such a range may be calculated for each color component and determined based on the magnitude.

  The achromatic color determination circuit 407 is a circuit that determines whether or not the image signal in the target section is substantially achromatic, and outputs the determination result of the achromatic color to the background portion detection circuit 410. The determination of achromaticity is realized, for example, by evaluating the color balance of the representative value described above. That is, the maximum value and the minimum value are extracted by mutual comparison of the representative value color components, the difference is obtained, and if the magnitude is smaller than a predetermined threshold as shown in Expression (7), it is determined that the color is achromatic. .

(Maximum representative value color component−minimum value of representative color component) <TH_N (7)
If so, achromatic. Here, TH_N is a constant.

  This achromatic determination result is also used to determine the background portion. The background targeted by this embodiment is white or an achromatic color similar to it, and the background such as colored paper is regarded as a color information portion that should not be corrected (removed) and is not handled as the background. For this reason, in order not to determine the color background portion as the background portion, the achromatic color determination circuit 407 determines whether or not the image signal is substantially achromatic.

  The brightness determination circuit 408 is a circuit that determines whether or not the image signal in the target section is bright, and outputs the brightness determination result to the background portion determination circuit 410. The determination of brightness can be realized, for example, by evaluating the brightness of each color component of the representative value. That is, as shown in Expression (8), if each color component of the representative value is brighter than a predetermined threshold value, it is determined to be bright.

  This brightness determination result is also used to determine the background portion. The background targeted by the present embodiment is a paper portion used for a document, and usually has a brightness as low as that of newspaper, and a dark portion such as black solid is not treated as a background. In order not to determine such a portion as the background, the brightness determination circuit 408 determines whether or not the image signal is bright.

  The light / dark change evaluation circuit 409 is a circuit that evaluates how the light / dark of the image signal in the target section changes compared to the surrounding image signals, and outputs the evaluation result of the light / dark change to the background determination circuit 410. To do. The evaluation of the light / dark change can be realized, for example, by referring to the predicted background level and the representative value. That is, the representative value is associated with the image signal in the attention section, and the predicted background level is associated with the surrounding image signals.

  The first light / dark change evaluation is an evaluation of whether or not the change in the dark direction is less than or equal to a predetermined amount compared to the surrounding image signals of the image signal of interest, for example, the predicted background level and the representative value Is calculated for each color component, and as shown in Expression (9), if each of them is smaller than a predetermined threshold, it is determined that the change in the dark direction is less than or equal to the predetermined amount.

  In addition, the second evaluation of the light / dark change is an evaluation of whether or not the change in the bright direction is greater than or equal to a predetermined amount by comparing the image signal in the target section with the surrounding image signal. For example, the representative value and the predicted background A difference from the level is calculated for each color component, and if any of them is larger than a predetermined threshold as shown in Expression (10), it is determined that the change in the bright direction is equal to or greater than a predetermined amount.

The background portion determination circuit 410 is a circuit that extracts a background level corresponding to the image signal of the target section, determines whether the image signal of the target section is the background portion or a portion equivalent thereto, and the determination result is sent to the background level holding circuit 412. While outputting, the background level extracted according to the determination result is output to the background level holding circuit 412 and the background level correction circuit 413.
These operations are performed according to, for example, the predicted background level, the representative value, the determination result of flatness, the determination result of achromaticity, the determination result of brightness, and the evaluation result of brightness change.

That is, if the evaluation result of flatness is flat, the determination result of achromatic color is achromatic, the determination result of brightness is bright, and the change result in the dark-to-dark change is less than a predetermined amount, the interval of interest Is determined as a true background portion, and a representative value is extracted as a background level.
In addition, although it is not a true background part, if the change in the bright direction is equal to or greater than a predetermined amount, the section of interest is determined as a quasi-background part, and a value calculated based on the predicted background level and the representative value is extracted. Set to the background level. For example, as shown in the equations (11) to (15), the extracted background level is calculated as follows: the maximum value of the representative value, and the maximum value obtained by calculating the difference between the representative value and the predicted background level for each color component, that is, The maximum amount of change in the bright direction is extracted, the predicted background level is corrected for each color component based on the maximum value, and the maximum value of the representative value is limited.

  Moreover, when it is neither a true background part nor a quasi-background part, it determines that the attention area is a non-background part. In this case, since the background level corresponding to the image signal in the target section cannot be extracted, the predicted background level is regarded as the extracted background level.

Maximum representative value = MAX (representative red component, representative green component, representative blue component) (11)
Maximum value of change in bright direction = MAX ((representative value red component−predicted background level red component), (representative value green component−predicted background level green component), (representative value blue component−predicted background level) Blue component)) (12)
Extracted background level red component = MAX ((predicted background level red component + maximum value of light direction change), (maximum representative value component)) (13)
Extracted background level green component = MAX ((predicted background level green component + maximum value of light direction change), (maximum representative value component)) (14)
Blue component of extracted background level = MAX ((blue component of predicted background level + maximum value of change in light direction), (maximum value component of representative value)) (15)
Here, in the expressions (11) to (15), MAX () is a function for selecting the maximum value of the elements. However, it is assumed that the predicted background level and the representative value have larger values as the brightness increases.

  Basically, the background portion determination circuit 410 may extract a representative value or a predicted background level based on whether or not the background portion is a true background portion, and even with this method, most manuscripts can be extracted without any problem. Can be handled.

  The background level holding circuit 412 is a circuit that holds the background level that represents the entire region in the main scanning direction and updates the representative background level according to the input image signal. Output to the background level correction circuit 413.

  The update of the representative background level is performed, for example, by the number of sections determined by the background determination circuit 410 as a true background and the total of the background levels extracted at that time, the number of sections determined as a quasi-background, and the extraction at that time. The total sum of the background levels is totaled for each main scan, and is performed based on this.

  That is, if there are a predetermined number (TH_T) or more of the sections determined to be true background portions, it is determined that the true background portions are sufficiently many and highly reliable values are obtained, and the sections determined to be true background portions. Based on the average value, the representative background level is updated for each color component as shown in Expression (16).

Renewed representative background level = (252 × retained representative background level + 4 × average value of sections determined to be true background) / 256 (16)
In addition, the number of sections determined to be true background portions is less than a predetermined number (TH_T), but if the number of sections determined to be quasi-background portions is greater than or equal to a predetermined number (TH_A), the number of quasi-background portions is sufficiently large and highly reliable. If the average value is brighter than the representative background level held, the representative background level is calculated for each color component based on the average value of the section determined to be a quasi-background background (24). Update like this.

  Renewed representative background level = (252 × retained representative background level + 4 × average value of sections determined to be quasi-background) / 256 (17)

  If the previous condition is not met, the representative background level is not updated.

In the above, the example in which the representative background level is updated based on the average value of all the sections determined as the true background portion or the quasi-background portion is shown, but the present invention is not limited to this. For example, it may be based on an average value when the number of sections determined to be true background portions reaches a predetermined number (TH_N). When a plurality of predetermined numbers are prepared and a larger predetermined number is reached Alternatively, the average value may be selected.
In addition, since the representative background level to be held is not fixed in the first main scanning (first line in the sub-scanning direction), a predetermined background level is held as an initial value instead.
Basically, the background portion holding circuit 412 counts only the total number of sections determined to be true background portions and the background level extracted at that time for each main scan, and updates the representative background level based on this. This method can handle most of the originals without any problem.

  The background level correction circuit 413 is a circuit that evaluates and corrects the reliability of the extracted background level, and outputs the corrected background level to the background level determination circuit 411. In the evaluation and correction of the extracted background level, for example, as shown in Expression (18), the reliability is evaluated by the magnitude of the difference from the representative background level for each color component, and when the reliability is low, that is, the difference is large. In this case, the allowable range is limited to the representative background level.

  The background level determination circuit 411 is a circuit that determines the background level of the target section based at least on the predicted background level and the corrected background level, and outputs the determined background level to the background level storage circuit 403 and the background level correction circuit 402. . The background level is determined by, for example, calculating a weighted average of the predicted background level and the extracted background level for each color component as shown in Expression (19), and setting this as the determined background level.

  Determined background level = ((corrected background level) +3 (predicted background level)) / 4 (19)

  The reason why the corrected background level is not used as the determined background level is that the calculation method of the extracted background level that is the basis of the corrected background level varies greatly depending on the background determination result. That is, if different background determinations are made in the surrounding sections, a large difference also occurs in the corrected background level, and this difference may appear as a pseudo contour after the image signal is corrected. That is, the equation (19) has an effect of reducing the influence of the difference by reducing the specific gravity of the corrected background level and suppressing the generation of the pseudo contour.

  The background level storage circuit 403 stores the determined background level and prepares for the output operation as described above. In the first main scanning (first line in the sub-scanning direction), there is no background level to be stored, but instead, the background level storage circuit 403 stores a predetermined background level as an initial value.

  A general manuscript has image information written on a blank sheet, and therefore, the image signal of the background portion has a feature that there are few bright and dark undulations, a bright value, and an achromatic color. For this reason, the background determination circuit 410 determines whether or not it is a true background by referring to the determination result of flatness, the determination result of achromaticity, the determination result of brightness, and the like. In particular, the character / thin line portion of the original and the highlight portion of the halftone dot printing are similar to the background portion in terms of achromaticity and brightness, so that the determination based on the evaluation result of the flatness is the background portion. This is effective for preventing erroneous determination.

  However, some manuscripts that are handled have continuous gradations such as silver halide photographs. These highlights have an achromatic color and brightness, but also have a flatness, and are erroneously determined as a background portion only under the above three conditions. For this reason, the background determination circuit 410 also refers to the first evaluation result of the change in brightness. That is, silver halide photographs and the like often have white borders or are affixed to a mount, and there are often subtle contrast differences between these boundaries. For this reason, the difference in these boundaries is determined by whether or not the change in the dark direction is less than or equal to a predetermined amount compared to the image signal in the area of interest compared to the surrounding image signals. Thereby, it is possible to prevent erroneous determination of a highlight portion of a silver halide photograph as a true background portion.

  If it is determined as a non-background portion, a predicted background level, that is, a background level that does not follow the image signal (representative value) of the target section is extracted, and this is reflected in the predicted background level in the next main scan or the like. The same result can be obtained in the processing after leaving.

On the other hand, as described above, if the change in the dark direction is equal to or less than the predetermined amount, the follow-up performance when the background level suddenly changes in the dark direction is deteriorated, which tends to cause a problem. In particular, when manuscripts with low whiteness paper such as newspaper are handled, the representative value extracted in the background portion becomes relatively dark. Therefore, if the predicted background level is not dark, the first change in light and darkness is caused. From the evaluation result, it is not determined to be a true background portion.
For this reason, the initial value stored in the background level storage circuit 403 in the first main scan needs to be a value that matches the lower limit brightness to be determined as the background, for example, the background level of newspaper.

  This value is preferably substantially the same as the evaluation criterion (TH_L) of the brightness determination circuit 408. That is, since the brightness determination circuit 408 also determines whether the image signal in the attention section has brightness that should be determined as the background portion, it is preferable to match the evaluation result.

Similarly, the initial value of the representative background level held by the background level holding circuit 412 in the first main scan is the same as the initial value of the background level storage circuit 403 or within the allowable range for the initial value of the representative background level. The initial value of the level storage circuit 403 needs to be entered.
This is because the background level correction circuit 413 does not work when the background portion determination circuit 410 performs non-background portion determination at the front end of the sub-scan and extracts the predicted background level, that is, the initial value of the background level storage circuit 403. Because. Thereby, it is possible to prevent the background level re-stored by the background level storage circuit 403 from being unnecessarily different from the initial value.

  In the background portion determination circuit 410, when determining the true background portion, the first evaluation result, the condition that “the change in the dark / dark change is less than a predetermined amount in the dark / dark change evaluation result” is added. In addition, a condition that “the change in brightness in the light / dark change evaluation result is a predetermined amount or less” is not added. This is because, in general, the brighter portion is more likely to be the background. However, it is also possible to assume that a document with a dark background level has punch holes. In this case, the portion of the document that is read through the punch hole is determined to be a true background, and the extracted background level is also the actual document level. It may become brighter than.

  The background level holding circuit 412 and the background level correction circuit 413 are effective in such a case. That is, since the background level holding circuit 412 holds the background level representing the main scanning direction, it is not easily affected by minute singular points such as punch holes, and the background level correction circuit 413 extracts the background level at the punch hole portions. The level can be evaluated as unreliable.

  By the way, in a manuscript having a halftone dot image on the entire surface, a true background portion is not determined due to its flatness. In addition, when a dark initial value is set in the background level storage circuit 403 as described above, there is a problem that the highlight portion is removed with the value as it is. Therefore, the background portion determination circuit 410 determines that the background portion is not a true background portion but is a quasi-background portion if the change in the bright direction is a predetermined amount or more, and calculates a value calculated based on the predicted background level and the representative value. The extracted background level is set. As a result, even in a document such as an entire halftone dot image, the background level is detected based on the image signal of the highlight portion, so that the highlight portion is not removed.

  The background level holding circuit 412 counts only the total number of sections determined to be quasi-background parts and the background level extracted at that time for each main scan, and the representative background level is updated more based on this. Because.

  In the above formulas (13) to (15), each color component of the predicted background level is corrected by a common value (maximum value of the change in the bright direction), and further limited by the maximum value component of the representative value. This is to eliminate the influence of the color of the representative value. As a result, a chromatic color portion such as a highlight color halftone dot portion can be made a semi-background without losing the color balance at the background level.

  The scanner unit 101 also reads a portion outside the original (such as a white original press plate) and outputs an image signal. Since the portion outside the document often has features similar to the background portion of the document, it is likely to be erroneously determined as a true background portion. On the other hand, in the present invention, the background level is predicted based at least on the background level of the section of interest. If the background level corresponding to the portion outside the document is referred to at the document edge, the background level is correctly set. It becomes unpredictable. For this reason, in the section located at the end of the document, or further outside, the original background level is not referred to as in the formula (20) or the formula (21) instead of referring to the background levels on both sides as illustrated in the formula (4). It is preferable to predict the background level without referring to the section in the outer direction.

Predicted background level on the left side of the document = (20 × (stored background level of the section of interest) + 12 × (stored background level of the right adjacent section)) / 32 (20)
Predicted background level on the right end side of the document = (20 × (stored background level of the target section) + 12 × (stored background level of the left end section)) / 32 (21)

  The digital color copying machine includes a document size sensor 212 and can specify a section located at the end of the document according to the detected size of the document. For this reason, the background level prediction circuit 404 also has a function of switching the background level prediction method according to the document size, thereby maintaining the accuracy of the document background level prediction at the document edge.

  The document size sensor is generally of a type that detects the size of a document of a standard size, and the document size sensor 212 of the present embodiment is also the same. For this reason, when a document of an irregular size is placed at a predetermined position on the platen 104, the document size sensor 212 detects the standard size having the most similar size characteristics as the size of the document. Therefore, in preparation for such a case, the background level prediction circuit 404 is caused to switch the background level prediction method with a size that is slightly smaller than the size detected by the document size sensor 212.

  In addition, as described above, in the present invention, the update of the background level representative of the entire region in the main scanning direction is performed by determining the number of sections determined as the true background portion, the sum of the background levels extracted at that time, and the quasi-background portion. The total number of sections and the background level extracted at that time are totalized during the main scan. Also in this case, if the portion outside the document is added to the total, it becomes difficult to obtain a correct representative background level. Therefore, the background level holding circuit 412 counts within a size range that is slightly smaller than the size detected by the document size sensor 212. It is preferable to carry out. Furthermore, the threshold value of the number of sections (TH_T, etc.) may be switched depending on the size.

  If the document size sensor 212 is not provided, it is preferable to set the background level prediction method switching and the total range in accordance with the assumed minimum document size.

  On the other hand, the background level correction circuit 402 receives the determined background level as the background level detected by the background level detection circuit 401, and corrects the input color image signal in accordance therewith. This correction is performed for each color component. Further, since the determined background level is input with a delay from the color image signal, the background level correction circuit 402 also performs processing corresponding to the delay.

An outline of the background correction processing performed by the background level correction circuit 402 will be described with reference to FIG.
FIG. 17 shows the input / output characteristics of the image signal in the background correction processing. Here, when correction according to the background level is not performed, a straight line having a slope of 1 is assumed. Further, the level that is the correction target of the background portion is shown as the reference background level.

Here, consider a case where the detected background level is brighter than the reference background level (A).
In this case, in an area where the input image signal is brighter than the detected background level, the input image signal having the same value as the detected background level is converted to the reference background level, and the image signal is converted by a straight line (A0-A1) having an inclination of 1. To do.

  In an area where the input image signal is darker than the detected background level, the input image signal having the same value as the detected background level is converted to the reference background level, and the image signal is converted with a predetermined inclination (A0-A2) smaller than 1. . However, in a region where this conversion result is larger than when correction is not performed, conversion without correction is performed.

Next, consider a case where the detected background level is darker than the reference background level (B).
In this case, in the region where the input image signal is brighter than the value (B ′) obtained by subtracting the predetermined threshold NR from the detected background level, the input image signal having the same value as the detected background level is converted to the reference background level, and the slope is 1. The image signal is converted by a certain straight line (B1-B0-B2).

  In an area where the input image signal is darker than B ', the image signal is converted by a straight line (B2-B3) having a predetermined inclination larger than 1 in B2. However, in a region where the conversion result is smaller than when no correction is performed, conversion without correction is performed.

In the above description, when the input image signal is larger than the detected background level, conversion is performed with a straight line having an inclination of 1 regardless of the brightness of the detected background level.
For example, this may be a conversion (A0-A1 ′) obtained by extending the straight line (A0-A2) or a conversion (B0-B1 ′) having the same inclination as the straight line (B2-B3). When changing across the reference background level, the conversion result of the image signal becomes discontinuous, which is not preferable. For this reason, when the input image signal is larger than the detected background level, this is prevented by performing conversion with a straight line having an inclination of 1.

  In the above description, the threshold NR is set, and the conversion gradient of the input image signal near the detected background level is set to 1. This may be converted by a straight line (B1'-B2 ') passing through B0 and having an inclination such as a straight line (B2-B3). However, in general, the image signal of the background portion is not necessarily uniform and has a certain degree of variation, and when converted by a straight line having an inclination such as a straight line (B1′−B2 ′), this variation is also enlarged. It may appear dirty. For this reason, by setting the threshold value NR and setting the conversion slope of the input image signal near the detected background level to 1, such an increase in variation is prevented.

  As described above, according to the present embodiment, it is possible to realize a background removal device that does not require pre-scanning, can follow two-dimensional fluctuations, and does not recognize a continuous tone highlight area as a background part. A document having a halftone dot image or the like on the entire surface can be handled without any problem.

[Second Embodiment]
The digital copying machine according to the present embodiment, which describes the second embodiment in which the present invention is suitably implemented, is different from the first embodiment in that it is a device that forms a monochrome image. In other words, the document unit for the electrophotographic process is provided with only black (K).

  FIG. 18 shows a configuration of a background removal apparatus when applied to an apparatus for forming a monochrome image. This background removing device is constituted by a background level detecting circuit 601 and a background level correcting circuit 602. The background level detection circuit 601 is a circuit that detects the background level of a document image based on an input image signal. The background level correction circuit 602 is a circuit that corrects the image signal in accordance with the detected background level.

The background level detection circuit 601 includes a background level storage circuit 603, a background level prediction circuit 604, a representative value extraction circuit 605, a flatness determination circuit 606, a brightness determination circuit 608, a light / dark change evaluation circuit 609, a background portion determination circuit 610, a background. A level determination circuit 611, a background level holding circuit 612, a background level correction circuit 613, and a sub-scanning position control circuit 617 are provided.
The background level storage circuit 603 is a circuit that stores the background level for each section when the main scanning direction of the image signal is divided into a plurality of sections. The stored background level is stored in the background level detection circuit 601. The input image signal is output to the background level prediction circuit 604 according to the position in the main scanning direction.

  The level prediction circuit 604 is based on at least the stored background level corresponding to the surrounding section of the section (target section) that is the processing target of the representative value extraction circuit 605 and the flatness determination circuit 606. The level prediction circuit outputs the predicted background level to the light / dark change evaluation circuit 609, the background portion determination circuit 610, and the background level determination circuit 611. The background level prediction method is the same as the background level prediction circuit 404 described above.

On the other hand, the input image signal is input to the representative value extraction circuit 605 and the flatness determination circuit 606.
The representative value extraction circuit 605 is a circuit that extracts and outputs an image signal value (representative value) that represents a corresponding section, that is, a section of interest, according to the position of the input image signal in the main scanning direction. The value is output to the brightness determination circuit 608, the light / dark change evaluation circuit 609, the background determination circuit 610, and the like. The representative value extraction method is the same as that of the representative value extraction circuit 405 described above.

  The flatness determination circuit 606 is a circuit that determines the flatness of the image signal in the attention section, and outputs the flatness determination result to the background portion determination circuit 610. The determination of flatness is realized by, for example, extracting the maximum value and the minimum value of the image signal in the target section, calculating the difference between them, and determining that the image signal is flat if each is smaller than a predetermined threshold value.

(Maximum value of image signal in attention interval−minimum value of image signal in attention interval) <TH_P
If so, flat. Here, TH_P is a constant. ... (31)

  The flatness determination result is used for determination of the background portion, as will be described later. That is, the image signals of the characters, fine line portions, and halftone dot printed portions of the document have a feature that the undulations of the light and dark are intense, and the flatness determination circuit 606 does not determine such a portion as a background portion. The presence or absence of undulations in the image signal is determined by flatness.

  Note that the flatness determination may be performed based on the image signal of the section of interest and its surroundings. Alternatively, a statistic such as the variance of the image signal in such a range may be calculated and determined based on the magnitude.

  The brightness determination circuit 608 is a circuit that determines whether or not the image signal in the target section is bright, and outputs the brightness determination result to the background determination circuit 610. The determination of brightness is realized by evaluating the brightness of the representative value. That is, if the representative value is brighter than a predetermined threshold value, it is determined that it is bright.

Typical value> TH_L '
Then it is bright. Here, TH_L ′ is a constant. However, the larger the representative value is, the larger the value will be. ... (32)

  The brightness determination result is also used to determine the background portion, as will be described later. The target background in this embodiment is a paper portion used for a document, and usually has a brightness of a newspaper level as a lower limit, and a dark portion such as a black solid is not treated as a background. In order not to determine such a portion as the background, the brightness determination circuit 608 determines whether the image signal is bright.

  The light / dark change evaluation circuit 609 is a circuit that evaluates how the lightness / darkness of the image signal in the section of interest changes compared to the surrounding image signals, and outputs the evaluation result of the light / dark change to the background determination circuit 610. To do. The evaluation of the change in brightness is performed with reference to the predicted background level and the representative value, for example. That is, the representative value is associated with the image signal in the attention section, and the predicted background level is associated with the surrounding image signals.

  The first light / dark change evaluation is an evaluation of whether or not the change in the dark direction is less than or equal to a predetermined amount compared to the surrounding image signals of the image signal of interest, for example, the predicted background level and the representative value And if each of them is smaller than a predetermined threshold, it is determined that the change in the dark direction is less than or equal to the predetermined amount.

(Predicted background level-representative value) <TH_D '
If so, the change in the dark direction is less than a predetermined amount.
Here, TH_D ′ is a constant. However, it is assumed that the predicted background level and the representative value are larger as the brightness is higher. ... (33)

  The second of the brightness change evaluation is an evaluation of whether or not the change in the bright direction is greater than or equal to a predetermined amount compared to the surrounding image signals in the target section. For example, the representative value and the predicted background level The difference is calculated, and if the difference is larger than a predetermined threshold, it is determined that the change in the bright direction is equal to or larger than the predetermined amount.

(Representative value-predicted background level)> TH_E '
Then, the change in the bright direction is more than a predetermined amount.
Here, TH_E ′ is a constant. However, it is assumed that the predicted background level and the representative value are larger as the brightness is higher. ... (34)

  The background portion determination circuit 610 is a circuit that extracts a background level corresponding to the image signal of the target section, determines whether the image signal of the target section is the background portion or a portion equivalent thereto, and determines the determination result as the background level holding circuit 612. And the background level extracted according to the determination result is output to the background level holding unit 612 and the background level correction unit 613. These operations are performed according to, for example, the predicted background level, the representative value, the flatness determination result, the brightness determination result, and the evaluation result of the brightness change.

That is, if the determination result of flatness is flat, the determination result of brightness is bright, and the change result in the light / dark change evaluation is less than or equal to a predetermined amount in the dark direction, the attention section is determined to be a true background portion, and the representative value Is the extracted background level.
If it is not a true background part but the change in the bright direction is greater than or equal to a predetermined amount, the section of interest is determined to be a quasi-background part, and the background level extracted in the same manner as described above.
When it is neither a true background part nor a quasi-background part, it is set as the background level which extracted the background level which judged the attention area as the non-background part and was estimated.

  Basically, the background portion determination circuit 610 may extract a representative value or a predicted background level depending on whether or not the background portion is a true background portion. Even with this method, most manuscripts can be handled without problems.

  The background level holding circuit 612 is a circuit that holds the background level representing the entire region in the main scanning direction and updates the representative background level according to the input image signal. Output to the level correction circuit 613. The update of the representative background level is extracted, for example, the number of sections determined by the background determination circuit 610 as the true background and the total of the background levels extracted at that time, the total number of sections determined as the quasi-background and the time. The sum total with the background level is totaled for each main scan and is performed based on this.

That is, if the number of sections determined to be true background portions is equal to or greater than a predetermined number (TH_T ′), it is determined that there are sufficiently many true background portions and a highly reliable value is obtained, and the true background is determined. The representative background level is updated as shown in Expression (16) based on the average value of the sections.
In addition, the number of sections determined to be true background is less than a predetermined number (TH_T ′), but if the number of sections determined to be quasi-skin is equal to or greater than a predetermined number (TH_A ′), the quasi-skin section is sufficient. If it is judged that many highly reliable values have been obtained, and if the average value is brighter than the representative background level held, the representative background level is calculated based on the average value of the section determined to be quasi-background. Update as in (17).
If these conditions are not satisfied, the representative background level is not updated.

  In addition, although the above description showed the example which updates a representative background level based on all the average values to the area determined to be a true background part or a quasi-background part, this invention is not limited to this. . For example, it may be based on an average value at the time when the number of sections determined to be true background portions reaches a predetermined number (TH_N ′). Furthermore, a plurality of predetermined numbers are prepared to reach a larger predetermined number. You may make it select the average value in the time of doing.

  In addition, since the representative background level to be held is not fixed in the first main scanning (first line in the sub-scanning direction), a predetermined background level is held as an initial value instead.

Basically, the background holding circuit 612 counts only the number of sections determined to be true background portions and the total number of background levels extracted at that time for each main scan,
Based on this, the representative background level may be updated, and even this method can handle most of the originals without any problem.

  The background level correction circuit 613 is a circuit that evaluates and corrects the reliability of the extracted background level, and outputs the corrected background level to the background level determination unit 611. Evaluation and correction of the extracted background level are performed, for example, as shown in Expression (18).

  The background level determination circuit 611 is a circuit that determines the background level of the target section based at least on the predicted background level and the corrected background level, and outputs the verified background level to the background level storage circuit 603 and the background level correction circuit 602. . The background level is determined, for example, as shown in Expression (19).

  The background level storage circuit 603 stores the determined background level and prepares for the output operation as described above. In the first main scanning (first line in the sub-scanning direction), there is no background level to be stored, but instead, a predetermined background level is stored as an initial value.

By the way, since the image information is recorded on a blank paper in a general document, the image signal of the background portion has a characteristic that it has a bright value with little shading of light and dark. Therefore, the background portion determination circuit 610 determines whether or not the background portion is a true background portion with reference to the determination result of flatness, the determination result of brightness, and the like. In particular, since the character / thin line of the original and the highlight portion of the halftone dot printing are similar to the background portion in terms of brightness, the evaluation result of the flatness is effective in preventing erroneous determination as the background portion.
However, some of the manuscripts that are handled have continuous gradations, such as silver halide photographs, and these highlights may have both brightness and flatness. There is a risk of misjudging it as a part. For this reason, the background portion determination circuit 610 also refers to the first evaluation result of the change in brightness. That is, many silver halide photographs have white borders or are affixed to a mount, and there are often subtle contrast differences at these boundaries. For this reason, the difference at these boundaries is determined by whether or not the change in the dark direction of the image signal in the target section is less than or equal to a predetermined amount compared to the surrounding image signals. Thereby, it can avoid misdetermining with a true background part. Note that a background level that is predicted to be determined as a non-background portion, that is, a background level that does not follow the image signal (representative value) of the target section, is extracted and reflected in the predicted background level in the next main scan. The same effect can be obtained in subsequent processing.

On the other hand, as described above, if the condition that the change in the dark direction is equal to or less than the predetermined amount is added as a condition, the follow-up performance in the case of a sudden change in the background level in the dark direction is likely to be a problem.
In particular, when dealing with a document with low whiteness such as newspaper, the representative value extracted in the background portion becomes relatively dark, and therefore the first evaluation of the change in light and darkness if the predicted background level is not dark. Depending on the result, it will not be determined as a true background. For this reason, the initial value stored in the background level storage unit 603 in the first main scan needs to be a value that matches the lower limit brightness (for example, the background level of newspaper) to be determined as the background. Note that this value is preferably substantially the same as the evaluation criterion (TH_L ′) of the brightness determination circuit 608 described above. That is, the brightness determination circuit 608 also determines whether or not the image signal in the attention section has brightness to be determined as the earlobe, and the evaluation results preferably match.

  Similarly, the initial value of the representative background level held by the background level holding circuit 612 in the first main scan is the same as the initial value of the background level storage circuit 603 or the background level is stored in an allowable range for the initial value of the representative background level. The initial value of the circuit 603 needs to be entered. This is because the background level correction circuit 613 does not work when the background portion determination circuit 610 performs non-background portion determination at the sub-scanning tip and extracts the predicted background level, that is, the initial value of the background level storage circuit 603. Therefore, it is possible to prevent the background level re-stored by the background level storage circuit 603 from being unnecessarily changed from the initial value.

In the background portion determination circuit 610, when determining the true background portion, a condition that the first evaluation result “brightness change evaluation result is less than or equal to a predetermined amount in the dark / dark change evaluation result” is added. In other words, the condition that “the change in the light / dark change evaluation result is less than a predetermined amount in the bright direction” is not added. This is because, generally, the brighter portion is more likely to be the background. However, it is also possible to assume that a document with a dark background has a punch hole. In this case, the document holding portion read through the punch hole is determined to be a true background, and the extracted background level is also higher than the actual document. May be brighter.
The background level holding circuit 612 and the background level correction circuit 613 are effective in such a case. That is, the background level holding circuit 612 holds the background level representative of the main scanning direction, and is therefore not easily influenced by minute singular points such as punch holes, and the background level correction circuit 613 extracts the background level from the punch holes. This is because the level can be evaluated as having low reliability.

  In a document having a halftone image on the entire surface, a true background portion is not determined due to the flatness. In addition, when a dark initial value is set in the background level storage circuit 603 as described above, there is a problem that a highlight portion is removed with the value as it is. Therefore, the background determination circuit 610 is not a true background, but is determined to be a quasi-background if the change in the bright direction is equal to or greater than a predetermined amount, and a value calculated based on the predicted background level and the representative value. The background level is extracted. As a result, even for a document such as a full screen halftone image, the background portion is extracted based on the image signal of the highlight portion, so that the highlight portion is not removed.

  Further, the background level holding circuit 612 counts only the total number of sections determined to be quasi-background and the background level extracted at that time for each main scan, and updates the representative background level based on this. This is also for the purpose of updating the representative background level more in a document having a halftone dot image on the entire surface.

  It should be noted that the correspondence to the reading of the portion outside the original (such as a white original pressure plate) is the same as that in the case of handling the color image signal.

  On the other hand, the background level correction circuit 602 receives the determined background level as the background level detected by the background level detection circuit 601 and corrects the input image signal accordingly. Since the determined background level is input with a delay from the image signal, the background level correction circuit 602 also performs processing corresponding to the delay. Note that the background correction processing performed by the background level correction circuit 602 is the same as that when a color image signal is handled.

  As described above, according to the present embodiment, the monochrome image forming apparatus including the background removal device that does not require pre-scanning, can follow the two-dimensional fluctuation, and does not recognize the continuous highlight area as the background. In particular, a document having a halftone image on the entire surface can be handled without any problem.

[Third Embodiment]
A third embodiment in which the present invention is preferably implemented will be described. FIG. 19 shows the configuration of the background removal circuit according to this embodiment.
The background removal circuit includes a background detection circuit 801 and a background correction circuit 802. The background detection circuit 801 detects the background level of the document based on the input image signal. The background correction circuit 802 corrects the image signal according to the background level detected by the background detection circuit 801.
The background detection circuit 801 includes a background level buffer (line buffer) circuit 803, a background level prediction circuit 804, a feature amount extraction circuit 805, and a background level determination circuit 806.
The background level buffer circuit 803 is a circuit (corresponding to the background level buffer in FIG. 9) that stores the background level of each section when the main scanning direction of the image signal is divided into a plurality of sections for each color component. The background level buffer circuit 803 outputs the stored background level to the background level prediction circuit 804 in accordance with the position of the image signal input to the background detection circuit 801 in the main scanning direction. As shown in FIG. 9, here, the main scanning direction of the image signal is divided with 16 pixels as one section.

The background level prediction circuit 804 is a previous line stored in the background level buffer circuit 803 as the background level of the section corresponding to the region (target region) focused on by the feature amount extraction circuit 805 and the left and right adjacent sections. This circuit predicts the background level of the region of interest based on the background level. The background level prediction circuit 804 outputs the predicted background level to the background level determination circuit 806. In the present embodiment, the background level is predicted by calculating the following weighted average.
Predicted background level = (2 × (background level of attention area corresponding section) + 7 × (background level of left adjacent section) + 7 × (background level of right adjacent section)) / 16 (22)

  FIG. 20 shows an example of the circuit configuration of the background level prediction circuit 804. The background level prediction circuit 804 includes flip / flop circuits (F / F circuits) 502, 504, and 506 and a product-sum circuit 509. A background level signal 503 output from the F / F circuit 502 in response to the input image signal 501 is input to the F / F circuit 504. The background level signal 505 output from the F / F circuit 504 is input to the F / F circuit 506. A section clock 508 corresponding to section switching is input to the F / F circuits 502, 504, and 506. The background level signals 503, 505, and 507 are a right adjacent section, an attention area corresponding section, and a left adjacent section, respectively. Corresponds to the background level.

  Background level signals 503, 505 and 507 output from each F / F circuit are input to a product-sum circuit 509. The product-sum circuit 509 performs the product-sum operation shown in the equation (22) for each color component, and outputs a predicted background level signal 510 representing the background level of the region of interest.

The feature amount extraction circuit 805 is a circuit that extracts the feature amount of the input document image signal.
The extracted feature amount is output to the background level determination circuit 806. In the present embodiment, the line image signal portion corresponding to one section of the background level buffer 803 is set as the attention area, and the feature amount extraction range (target range) is also set as the attention area. The maximum value, the minimum value, and the average value are obtained for each signal component and output as a feature value.

FIG. 21 shows a more detailed circuit configuration of the feature quantity extraction circuit 805. In addition, what is illustrated is a configuration corresponding to one signal component.
The image signal input to the feature amount extraction circuit 805 is input to the comparators 652 and 653, the selectors 654 and 655, and the adder 656. Output signals of the F / F circuits 659 and 660 are input to the other input terminals of the comparators 652 and 653 and the selectors 654 and 655 via the selectors 657 and 658. Then, the comparator 652 compares the magnitudes of both inputs, and outputs a control signal 661 to the selector 654 so that the selector 604 selects the smaller signal. The comparator 653 compares the magnitudes of both inputs, and outputs a control signal 662 to the selector 655 so that the selector 655 selects the larger signal.

  Output signals from the selectors 654 and 655 are input to the F / F circuits 659 and 660, respectively. F / F circuits 659 and 660 are provided. Each selection result is held in accordance with the synchronizing signal 663 of the image signal.

  Further, the maximum value and the minimum value that can be taken by the image signal are input to the other input terminals of the selectors 657 and 658, respectively. The operations of the selectors 657 and 658 are controlled by the section clock 508. That is, the maximum value and the minimum value are selected for switching between sections.

  The respective output signals of the F / F circuits 659 and 660 are also input to the F / F circuits 664 and 665, and the F / F circuits 664 and 665 respond to the section clock 508 in accordance with the F / F circuits 659 and 660. Each output signal is held.

  Thereby, the maximum value and the minimum value of the image signal in the target range are output from the F / F circuits 664 and 665, respectively.

  On the other hand, the output signal of the adder 656 is input to the F / F circuit 666. The F / F circuit 616 holds this in accordance with the synchronization signal 663 of the image signal.

  The output signal of the F / F circuit 666 is connected to the other input terminal of the adder 656 via the selector 667. The selector 667 is controlled by the section clock 508, “20” is input to the other input terminal, and “0” is selected when the section is switched. The output signal of the F / F circuit 666 is also input to the F / F circuit 668. The F / F circuit 668 holds the output signal of the F / F circuit 666 according to the section clock 508. As a result, the sum of the image signals in the target range is output from the F / F circuit 668. As described above, in this embodiment, since one section is 16 pixels, the value of 1/16 of the output signal of the F / F circuit 668 indicating the total sum of the image signals in the target range, that is, the lower 4 bits The value excluding is the average value of the image signal in the target range.

  The background level determination circuit 806 is a circuit that determines the background level of the region of interest based on the predicted background level and the extraction amount. The determined background level is output and stored in the background level buffer circuit 803, and is output to the background correction circuit 802 as a detection result of the background level of the region of interest.

Here, the background level determination circuit 806 for obtaining the background level used for background correction will be described in detail.
In the background level determination procedure, whether the image of the attention area is at the background level is determined based on the feature amount of the input image, but the background portion of the document is usually the paper itself used for the document. Generally, it has the characteristics that there are few undulations of light and dark, achromatic color, and brighter than a predetermined brightness. Therefore, the background determination of the document is performed based on these three characteristics.

Since the maximum value, the minimum value, and the average value of each component of the image signal of the target region are input from the feature amount extraction circuit 805, the background level determination circuit 806 can convert the above three features by the following method. evaluate.
○ To determine whether there are few undulations in light and dark, the difference between the maximum value and the minimum value is obtained for each component and each is compared with a predetermined threshold value.
○ To determine whether the color is achromatic, the difference between the maximum value and the minimum value of the average value of each component is obtained and compared with a predetermined threshold value.
○ To determine whether the brightness is higher than a predetermined brightness, an average value of each component is obtained and compared with a predetermined threshold value.

If these evaluation results include all the features of the background portion, the background level determination circuit 806 determines the target region as the background portion, and predicts the predicted background level and the target region as shown in Expression (23) for each component. The background level is determined by weighted averaging of the average value of the image signals at.
In this example, the weight of the predicted background level is increased, and the weight of the target area is relatively decreased, so that the predicted background level is mainly used as a base, and this is corrected by the image of the target range.
Determined background level = (1 × (average value of target range) + 7 × (predicted background level)) / 8 (23)

  On the other hand, in this example, if there is a feature evaluation result that does not include the feature of the background portion, the target region is not determined as the background portion. In this case, since there is no basis for correcting the predicted background level, the predicted background level is determined as it is as the background level.

A more detailed circuit configuration of the background level determination circuit 806 will be described. FIG. 22 is a diagram illustrating a circuit configuration of the background determination unit of the target region.
In FIG. 22, each of MaXR, MaxG, MaxB, MinR, MinG, MinB, AveR, AveG, and AveB represents the maximum value, minimum value, and average value of the image signal in the target range output from the feature amount extraction circuit 805. It is a component signal. The MaxR and MinR signals, the MaxG and MinG signals, and the MaxB and MinB signals are input to the difference circuits 701, 702, and 703, respectively. Difference circuits 701, 702, and 703 are circuits that output the difference between the input signals, and output the results to comparators 704, 705, and 706, respectively.

  Predetermined threshold signals DifTHr, DifTHg, and DifTHb are input to the comparators 704, 705, and 706, respectively. Comparators 704, 705, and 706 output true to AND gate 707 if the difference values are smaller than threshold values DifTHr, DifTHg, and DifTHb, respectively. As a result, the AND gate 707 outputs true when there are few bright and dark undulations.

  The AveR, AveG, and AveB signals are input to comparators 708, 709, and 710 together with predetermined threshold signals DarkTHr, DarkTHg, and DarkTHb, respectively. Each of the comparators 708, 709, and 710 outputs true to the AND gate 711 if the average value is larger than the threshold value. As a result, the AND gate 711 outputs true when the brightness is above a certain level.

  The AveR, AveG, and AveB signals are also input to the maximum value selection circuit 712 and the minimum value selection circuit 713, respectively. The maximum value selection circuit 712 selects the maximum value of the average value of each color component and outputs it to the difference circuit 714. Also, the minimum value selection circuit 713 selects the minimum value of the average value of each color component and outputs it to the difference circuit 714. The difference circuit 714 is a circuit that outputs a difference between input signals, and inputs the result to the comparator 715.

  The comparator 715 also receives the threshold signal BalTH, and outputs true if the difference value is smaller than the threshold value. Thereby, true is output when the image signal of the target region is achromatic.

  The outputs of the AND gates 707 and 711 and the comparator 715 are input to the AND gate 716. The AND gate 716 outputs true when the target area has all the features of the background portion.

  FIG. 23 shows the configuration of the determination unit that receives the background determination result and determines the background level. ExpR, ExpG, and ExpB are predicted background level signals output from the background level prediction circuit 804. The ExpR, ExpG, and ExpB signals are input to the difference circuits 721, 722, and 723, together with the average values AveR, AveG, and AveB of the image signals in the target region. The difference circuits 721, 722, and 723 are circuits that output the difference between the input signals, and the results are input to the comparators 724, 725, and 726, respectively.

  The comparators 724, 725, and 726 also receive predetermined threshold values DownTHr, DownThg, and DownTHb, respectively. Comparators 724, 725, and 726 each output true to AND gate 727 if the difference is less than the threshold. Here, the difference circuit outputs the sex value when the average value is darker than the predicted background level. Accordingly, the AND gate 727 outputs true when the average value of each color component is slightly darker or brighter than the predicted background level.

Output signals of the AND gate 716 and the AND gate 727 are input to the AND gate 728, respectively.
On the other hand, the weighted average circuits 729, 730, and 731 perform the correction calculation of the predicted background level shown in the above equation (23). The weighted average circuits 729, 730, and 731 are input with the predicted background level signals ExpR, ExpG, and ExpB and the average values AveR, AveG, and AveB of the image signals in the target region. The weighted average circuits 729, 730, and 731 output the calculation results to the selectors 732, 733, and 734, respectively. The other input terminals of the selectors 732, 733, and 734 receive ExpR, ExpG, and ExpB signals, respectively, and their operations are controlled by the output signal of the AND gate 728.

  The output signal of the AND gate 728 is true when the target area has all the features of the background portion and the average value of each color component is slightly darker or brighter than the predicted background level. At this time, the selectors 732, 733, and 734 select and output the calculation result of the weighted average circuit. Note that the latter condition, that is, the condition of the AND gate 727, prevents the background level from following the background by determining that the highlight portion of the image whose density changes continuously as in the case of a silver salt photograph is the background. It is particularly effective. That is, it is possible to identify the difference in density between the highlight portion and the true background portion based on whether or not the average is slightly darker than the predicted background level, thereby preventing the background level from following.

  As described above, in the present embodiment, the background level determined by the background level determination circuit 806 is output from the selectors 732, 733, and 734.

Next, background correction performed in response to the background level determined by the background level determination circuit 806 will be described.
The background correction circuit 802 receives the background level detected by the background detection circuit 801 (the background level determined by the background level determination circuit 806) and also receives an image signal via the delay buffer 807. The background correction circuit 802 corrects the input image signal according to the background level.

The input / output characteristics of the image data in the background correction processing in this embodiment are the same as those shown in FIG. FIG. 17 shows a comparison between the case where the background correction is performed and the case where the background correction is not performed. When correction according to the background level is not performed, a straight line with an inclination of 1 is obtained. In addition, a predetermined output level that is a correction target for the background portion is shown as a reference background level.
The background level correction method according to the present embodiment changes the conversion characteristics provided to the image signal of the target region in accordance with the background level detected by the background detection circuit 801. Since this is the same as the background correction in the first embodiment, a duplicate description is omitted.

A circuit configuration of the background correction circuit 802 will be described. FIG. 24 shows a circuit configuration of the background correction circuit 802.
The detected background level signal Detect input from the background detection circuit 801 is input to the comparator 951 together with the reference background level signal BaseW corresponding to the reference background level. Thereby, it is determined whether or not the detected background level is darker than the reference background level.

The selector 952, the subtractor 953, and the comparator 954 are circuits that determine whether the input / output characteristics having the slope 1 are used for the background correction processing, for example, the characteristic line (B1-B0-B2) or characteristics in the example of FIG. A circuit for determining whether or not to perform conversion by the line (A0-A1) is configured.
A signal NR corresponding to the threshold NR or a fixed value “0” is input to the selector 952, and selection is performed based on the output signal of the comparator 951. That is, the selector 952 outputs the threshold signal NR when darker than the reference background level, and outputs “0” when brighter than the reference background level.
The subtractor 953 subtracts the output signal of the selector 952 from the detected background level signal Detect. Thereby, the value of B2 or A0 in FIG. 17 is obtained.
The comparator 954 compares the value B2 or A0 input from the subtractor 953 with the image signal ImgIN input from the delay buffer 807, and compares the characteristic line (B1-B0-B2) or the characteristic line (A0-A1). ) To determine whether or not to perform conversion.

  The subtractor 955 and the adder 956 constitute a circuit that performs correction when the determination output of the comparator 954 is true. That is, the subtractor 955 calculates the difference between the reference background level signal BaseW and the detected background level Detect to obtain the correction amount, and adds the subtraction result to the image signal ImgIN to obtain the corrected image signal ImgOUTA. .

  Next, a case where input / output conversion is performed with characteristics where the slope is not 1 will be described. In FIG. 17, input / output conversion is performed with the characteristic (B2-B3) or (A0-A2), and conversion is performed using B2 or A0 as a base point.

  The selector 957 selects the tilting unit of the characteristic line (B2-B3) or the characteristic line (A0-A2) portion, and sets the tilts to “7/4” and “3/4”, respectively, and compares them with the comparator 951. Output signal (a signal indicating whether or not the detected background level is darker than the reference background level). Further, a difference between B2 or A0 and the image signal ImgIN is obtained by a subtracter 958, and a slope selected by the selector 957 is multiplied by this by a multiplier 959. In this way, the corrected difference with B2 or A0 as the base point can be obtained.

  On the other hand, the corrected value of B2 or A0 is obtained by subtracting the output signal of the selector 952 from the reference background level signal BaseW by the subtractor 960. Accordingly, the subtracter 961 subtracts the output of the multiplier 959 from the output of the subtractor 960, and the image signal ImgOUTB corrected by the characteristic line (B2-B3) or the characteristic line (A0-A2) is obtained.

  However, since the image signal ImgOUTB also includes conversion on the characteristic line (B2-B3) or an extension line of the characteristic line (A0-A2), the result is compared with a case where the result is not corrected by the comparator 962. Thus, the image signal ImgOUTB or the image signal ImgIN (when correction is not performed) is selected by the selector 963, and the selector 963 outputs the image signal ImgOUTB ′.

  The selectors 964 and 965 select the image signal ImgOUTB or the image signal ImgIN, respectively, according to the output signal of the comparator 951 (a signal indicating whether or not the detected background level is darker than the reference background level). Accordingly, whether the image signal ImgOUTB or the image signal ImgIN is selected is reversed depending on whether or not the detected background level is darker than the reference background level.

  Further, the image signal ImgOUTA or ImgOUTB ′ obtained as described above is selected by the selector 966 controlled by the output signal of the comparator 954 (a signal indicating whether or not the conversion is performed with the input / output characteristics of the slope 1). The image signal ImgOUT corrected according to the detected background level signal Detect is obtained.

In the background correction processing method according to the present invention, the functional blocks shown in FIGS. 2A and 2B may be configured on a computer by a program.
The program (background correction processing program according to the present invention) is a computer-readable information recording medium (CD-ROM, flexible disk, DVD) as a file that can be installed on a computer or a computer-executable file. Or may be downloaded to a computer via a network.

  Each of the above embodiments is an example of a preferred embodiment of the present invention, and the present invention is not limited to this and can be variously modified.

It is a figure which shows an example of the input-output conversion characteristic of the image signal by this invention. It is a figure which shows the structure of a background correction | amendment processing apparatus. It is a figure which shows another example of the input-output conversion characteristic of the image signal by this invention. It is a figure which shows another example of the input-output conversion characteristic of the image signal by this invention. It is a figure which shows another example of the input-output conversion characteristic of the image signal by this invention. It is a figure which shows another example of the input-output conversion characteristic of the image signal by this invention. It is a figure which shows another example of the input-output conversion characteristic of the image signal by this invention. It is a figure which shows another example of the input-output conversion characteristic of the image signal by this invention. It is a figure for demonstrating the outline of a background correction process. 1 is a diagram showing the structure of a mechanism system of a color image copying apparatus according to the present invention. 1 is a diagram showing a configuration of a transmission system of a color image copying apparatus according to the present invention. It is a figure which shows the structure of a scanner unit. It is a figure which shows the structure of an image processing unit. FIG. 3 is a diagram illustrating a configuration of a printer unit. It is a figure which shows the structure of an operation display unit. It is a figure which shows the structure of the background removal circuit concerning 1st Embodiment which implemented this invention suitably. It is a figure which shows an example of the input-output conversion characteristic of the image signal in this invention. It is a figure which shows the structure of the background removal circuit concerning 2nd Embodiment which implemented this invention suitably. It is a figure which shows the structure of the background removal circuit concerning 3rd Embodiment which implemented this invention suitably. It is a figure which shows the circuit structure of a background level prediction circuit. It is a figure which shows the circuit structure of a background level determination circuit. It is a figure which shows the circuit structure of a background determination part. It is a figure which shows the circuit structure of a determination part. It is a figure which shows the circuit structure of a background correction circuit. It is a figure which shows an example of the input-output conversion characteristic of the image signal by a prior art.

Explanation of symbols

101 Scanner Unit 102 Printer Unit 103, 301 Document 104 Contact Glass 105 Halogen Lamp 106 First Mirror 107 Second Mirror 108 Third Mirror 109 Lens 110 CCD
111 Image Processing Unit 112 Polygon Mirror 113 fθ Lens 114 Fourth Mirror 115 Photosensitive Drum 116 Polygon Motor 117 Charging Charger 118 Document Unit 119 Transfer Belt 120a, 120b Paper Feed Cassette 121a, 121b Recording Paper 122a, 122b Paper Feed Roller 123a, 123b Registration roller 124 Transfer charger 125a, 125b Thermal fixing unit 126 Cleaning unit 201 Operation display unit 202 Digital copying machine 203 System control unit 204 LAN connection device 205 LAN
206 workstation 207 information processing terminal 208 A / D conversion circuit 209 shading correction circuit 210 scanner control circuit 211 carriage drive motor 212 document size sensor 213 background removal circuit 214 gamma conversion circuit 215 image area separation circuit 216 area control circuit 217 delay circuit 218 Filter circuit 219 Color correction circuit 220 Scaling circuit 221 Gradation processing circuit 222 Image processing control circuit 223 LD control circuit 224 LD
225 Printer control circuit 226, 227 Motor 228 High voltage power supply 229 Paper size sensor 230 TPD (touch panel display)
231 Keyboard 232 Operation display control circuit 302 Image data range 303 Background level buffer 304 Sub-scanning position 305 Section 306 Background level prediction unit 307 Feature amount extraction unit 308 Background level determination processing unit 401, 601 Background level detection circuit 402, 602 Background level correction Circuits 403, 603 Background level storage circuit 404, 604 Background level prediction circuit 405, 605 Representative value extraction circuit 406, 606 Flatness determination circuit 407 Achromaticity determination circuit 408, 608 Brightness determination circuit 409, 609 Light / dark change evaluation circuit 410 , 610 Background level determination circuit 411, 611 Background level determination circuit 412, 612 Background level holding circuit 413, 613 Background level correction circuit 417 Sub-scanning position control circuit 501 Input image signal 502, 504, 506, 6 9,660,664,665,666 flip-flop circuit 503, 505, 507 background level signal 508 interval clock 509 product-sum circuit 510 predicted background level signal 651
652, 653, 954 Comparator 654, 655, 657, 658, 667, 952, 957, 963, 964, 965, 966 Selector 656, 956 Adder 661, 662 Control signal 663 Sync signal 701, 702, 703, 714 Difference Circuits 704, 705, 706, 708, 709, 710, 715, 951, 962 Comparator 707, 711, 716 AND gate 712 Maximum value selection circuit 713 Minimum value selection circuit 901, 911 Image signal generator 902, 912 Background detection processing Devices 903, 913 Background correction processing device 904 Arithmetic unit 905 Conversion unit 906, 916 Image signal receiving device 907, 917 System controller 953, 955, 958, 960, 961 Subtractor 959 Multiplier

Claims (13)

When the intensity of the image signal of the background portion of the target image is equal to the intensity of the image signal of the reference background level, the input image data is output in an uncorrected equivalent state, and the intensity of the image signal of the background portion of the target image is When the intensity of the image signal at the reference background level is weaker, the image signal of the background portion of the target image is corrected to the reference background level, and the substantially black image signal is output in an uncorrected equivalent state. A background correction processing apparatus that
The intensity of the image signal at the reference background level is A, the intensity of the black image signal is B, the intensity of the image signal of the background portion of the target image is C, the input image data is x, and the minute displacement of the input image data Is Δ and the output data after the background correction processing is y,
The image data input / output conversion function y = ∫ (x) when the background level C of the target image is weaker than the reference background level A is near the intensity C of the image signal of the background portion of the target image.
(AB) / (CB)> (∫ (C + Δ) −∫ (C)) / Δ> 0
And (A−B) / (C−B)> (∫ (C) −∫ (C−Δ))> 0
A background correction processing apparatus having the following characteristics: The image data input / output conversion function y = ∫ (x) is
1 = (∫ (c + Δ) −∫ (C)) / Δ
And 1 = (∫ (C) −∫ (C−Δ)) / Δ
The background correction processing apparatus according to claim 1, wherein: The image data input / output conversion function y = ∫ (x) is
1> (∫ (c + Δ) −∫ (C)) / Δ
And 1> (∫ (C) −∫ (C−Δ)) / Δ
The background correction processing apparatus according to claim 1, wherein: When the intensity of the image signal of the background portion of the target image is equal to the intensity of the image signal of the reference background level, the input image data is output in an uncorrected equivalent state, and the intensity of the image signal of the background portion of the target image is When the intensity of the image signal at the reference background level is weaker, the image signal of the background portion of the target image is corrected to the reference background level, and the substantially black image signal is output in an uncorrected equivalent state. A background correction processing method,
The intensity of the image signal at the reference background level is A, the intensity of the black image signal is B, the intensity of the image signal of the background portion of the target image is C, the input image data is x, and the minute displacement of the input image data Is Δ and the output data after the background correction processing is y,
The image data input / output conversion function y = ∫ (x) when the background level C of the target image is weaker than the reference background level A is near the intensity C of the image signal of the background portion of the target image.
(AB) / (CB)> (∫ (C + Δ) −∫ (C)) / Δ> 0
And (A−B) / (C−B)> (∫ (C) −∫ (C−Δ)) / Δ> 0
The background correction processing method characterized by having the following characteristic. The image data input / output conversion function y = ∫ (x) is
1 = (∫ (c + Δ) −∫ (C)) / Δ
And 1 = (∫ (C) −∫ (C−Δ)) / Δ
The background correction processing method according to claim 4, wherein: The image data input / output conversion function y = ∫ (x) is
1> (∫ (c + Δ) −∫ (C)) / Δ
And 1> (∫ (C) −∫ (C−Δ)) / Δ
The background correction processing method according to claim 4, wherein:   The program which makes a substantial computer perform the background correction | amendment processing method of any one of Claim 4 to 6. Based on image data obtained by main / sub-scanning the target image, first background detection means for detecting a representative background level that is a background level representing the entire image, and based on the detected representative background level An image processing apparatus comprising: a second background detection unit that detects a local background level that is a local background level of the target image; and a background correction unit that corrects the image data based on the local background level. Because
The second background detection means includes
Storage means for storing the background level in each section when the main scanning direction of the image data is divided into a plurality of sections;
A background level predicting means for predicting the background level of the target section based on the background level around the target section stored in the storage means when any one of the sections is the target section;
Extraction means for extracting only the attention section or the attention section and its peripheral section as a target section, and extracting the feature amount of the image data of the target section;
Determining means for determining the local background level based on the predicted background level and the background level;
The image processing apparatus, wherein the storage unit stores the representative background level as an initial value, and stores the local background level after the start of the main / sub-scan.   The image processing apparatus according to claim 8, wherein the prediction unit predicts a background level by performing a weighted average of a background level of a section around the section of interest stored in the storage unit. The extraction means includes means for extracting the undulations of light and dark as the feature amount of the image data of the target section,
The determination means is a background determination means for determining whether or not the image of the target section is regarded as a background by threshold processing for light and dark undulations extracted by the extraction means;
10. The image processing apparatus according to claim 8, further comprising: a unit that determines a background level predicted by the prediction unit as the local background level when the background determination unit determines that the background is not regarded as a background. 11. The extraction means includes means for extracting saturation as a feature amount of the image data of the target section,
The determining means determines whether or not to consider that the image of the target section is the background according to the saturation extracted by the extracting means;
11. The apparatus according to claim 8, further comprising: a unit that determines the background level predicted by the prediction unit as the local background level when the background determination unit determines that the background is not regarded as the background. Image processing apparatus. The extraction means includes means for extracting an average brightness as a feature amount of the image data of the target section,
The determination means determines whether or not the image of the target section is considered to be background by threshold processing for the average brightness extracted by the extraction means;
12. The apparatus according to claim 8, further comprising: a unit that determines the background level predicted by the prediction unit as the local background level when the background determination unit determines that the background is not regarded as the background. Image processing apparatus. The extraction means includes image average value calculation means for calculating an average value of the image data of the target section as a feature amount of the image data of the target section,
The determining means includes a weighted average background level calculating means for calculating a weighted average of the background level predicted by the predicting means and the average value calculated by the image average value calculating means;
13. The apparatus according to claim 10, further comprising: a unit that determines the background level calculated by the background level calculation unit as the local background level when the difference determination unit determines that the background can be regarded as a background. An image processing apparatus according to claim 1.

Images